KR19980033190A - A microwave apparatus for cleaning an in-situ vacuum line of a substrate processing apparatus - Google Patents

Abstract

The present invention relates to an apparatus for minimizing impurity deposition of a discharge line of a substrate processing chamber. The device according to the invention comprises first and second electrodes having opposed surfaces defining a fluid conduit therebetween. The fluid conduit includes an inlet, an outlet, and a collection chamber between the inlet and the outlet. An apparatus according to the present invention is connected to an inlet of the substrate processing chamber for receiving an effluent of the substrate processing chamber. The collection chamber is configured and arranged to collect suspended solids flowing through the fluid conduit to inhibit the discharge of the suspended solids. The microwave plasma generation system supplies microwave energy into the fluid conduit to form a plasma from the etchant gas in the fluid conduit. The components from the plasma react with the suspended solids collected in the collection chamber to form a gaseous product that can be pumped from the fluid conduit. The apparatus according to the present invention may further comprise an electrostatic collector to strengthen the collected particles in the collection chamber to inhibit the discharge of the suspended solids.

Description

A microwave apparatus for cleaning an in-situ vacuum line of a substrate processing apparatus

The present invention relates generally to the field of semiconductor processing equipment, and in particular to a method and apparatus for removing impurities and residues from the interior of a vacuum discharge line connected to a process chamber and to a process for reducing perfluorocompound (PFC) gas emissions from the process chamber ≪ / RTI >

During the chemical vapor deposition (CVD) process, the deposition gas is released into the processing chamber to form a thin film layer on the surface of the processed substrate. During the CVD process unwanted deposition occurs on regions such as the walls of the processing chamber. However, because the time each individual particle of these deposition gases resides in the chamber is very short, only a portion of the particles that are released into the chamber are consumed during the deposition process and deposited on one of the wafer or chamber walls.

Unspent gas particles are pumped from the chamber along with partially reacted compounds and reaction byproducts through a vacuum line, commonly referred to as the foreline. Many of these released gases include residues or suspended solids which are in a highly reactive state and / or which may form undesirable deposits in the foreline. Deposition and accumulation of flour residue and / or suspended solids at a given time causes problems. First, the solids are spontaneous ignitables that cause problems when the vacuum seal is broken and the foreline is exposed to ambient conditions during a periodic standard cleaning operation. Second, when a large amount of deposition material accumulates in the foreline, the foreline and / or the vacuum pump associated with it are clogged when not properly cleaned. Even during periodic cleaning, the accumulation of solids hinders normal operation of the vacuum pump and shortens the life of the pump. In addition, the solid material can countercurrent flow from the foreline to the process chamber and contaminates processing steps that adversely affect wafer yield.

To overcome these problems, the inner surface of the foreline must be cleaned regularly to remove the deposition material. This process is performed during a standard chamber cleaning operation used to remove unwanted deposition material from the chamber walls and similar areas of the processing chamber. Common chamber cleaning techniques include the use of etching gases such as fluorine to remove material deposited from the chamber walls and other areas. This etching gas is introduced into the chamber, and the plasma is formed so that the etching gas reacts with the deposition material from the chamber walls to remove the deposited material. This cleaning process is generally performed between deposition steps for all wafers or all N wafers.

Removing the deposition material from the chamber walls is very simple in that the plasma occurs in the region in the plasma proximate to the deposition material. Removing the deposition material from the foreline is more complicated because the foreline is downstream of the processing chamber. In a fixed time period, most points in the processing chamber are in further contact with the cantfluoro atoms at the points in the foreline. Thus, in a fixed time period, the chamber can be sufficiently cleaned by a cleaning process while remaining and similar deposits remain in the foreline.

To sufficiently clean the foreline, the cycle of the cleaning operation must be increased. However, increasing the length of the cleaning operation is undesirable because it adversely affects wafer throughput. This residue accumulation can also be washed to such an extent that the reactants from the washing step can be released into the foreline in a state where the reactants can react with the residue in the foreline. In some systems and applications, the lifetime of the discharged reactant is not sufficient to reach even the end or middle portion of the foreline. In these systems and applications, residue accumulation is an important concern. Thus, there is a need for an apparatus for effectively and thoroughly cleaning a foreline in a semiconductor processing system and a method of performing the same.

One method that has been used to clean the foreline is a cleaning system that uses plasma enhanced CVD technology to extract reactive components from the offgas as a film deposition on the electrode surface. The cleaning system is designed as a solid film to maximize the removal of reactants and uses a spiral electrode with a large surface area. The helical electrode is contained within a movable cross-shaped vacuum cleaner located near the end of the blower pump and the mechanical pump cyborg line. After a large amount of solid contaminants have accumulated on the electrode, the cross-shaped electric vacuum cleaner can be removed for processing and replacement.

Conventional methods that follow a large surface area of the electrode that provides a region for the solid material deposited by the system for collection have problems. In order to accommodate the large surface area of the electrode, the system must be essentially large. Moreover, conventional cleaning systems cause high costs because removable cross-shaped electric vacuum cleaners are disposable products that must be replaced properly. Also, the cleaning system can not reliably remove the flour or flocculent solids that are located downstream of the beginning of the vacuum foreline and accumulate at the beginning of the foreline.

Thus, there is a need for improved methods and apparatus for cleaning foreline.

Another important problem in CVD and other substrate processing apparatuses relates to the form of gases and by-products discharged from the process chamber through the front side. For example, because the separation of gases in the cleaning plasma is incomplete (only 10% of the gas particles introduced in some applications are separated) and the time each of the particles of the cleaning gas resides in the chamber is very short, Only a small fraction of the particles react with the deposited material. Gas particles that do not partially react with the etchant are pumped from the chamber along with the etchant and reactive by-products through a vacuum line, referred to as the so-called foreline. The discharged gas is a by-product generated in the semiconductor process.

Many fluoro containing gases used in the semiconductor industry as cleaning etchant gases are referred to as perfluoro compounds or simply as PFC. Some of the more commonly used PFCs include CF ₄ , C ₂ F ₆ , NF _3, and SF ₆ or similar gases. These gases are known to have a long lifetime (eg, CF ₄ = 50,000 years) and cause global warming effects. Therefore, discharging them to the atmosphere is very harmful and should be regulated. Therefore, it is important to reduce the PFC emitted from the semiconductor processing apparatus such as the CVD reaction chamber.

It is an object of the present invention to solve the problems of the prior art by providing an apparatus for preventing floating solids and other residues from accumulating in the discharge line of a substrate processing apparatus and reducing PFC emission from the substrate processing apparatus.

It is another object of the present invention to provide a method and apparatus optimized for either particle reduction or PFC emission reduction.

It is a further object of the present invention to provide a method and apparatus for both particle and PFC emission reduction for use in any substrate processing operation.

BRIEF DESCRIPTION OF THE DRAWINGS Figure 1 shows one embodiment of a simplified chemical vapor deposition apparatus to which the apparatus of the present invention may be attached.

2 is a view showing a first method of connecting the present invention to the chemical vapor deposition apparatus of FIG.

3 shows a second method of connecting the present invention to the chemical vapor deposition apparatus of Fig.

Figure 4 (a) shows a preferred embodiment of the apparatus of the present invention in which particles can be optimally reduced without vacuum (vacuum line cleaning).

Fig. 4 (b) is a front view of the vacuum line cleaning apparatus shown in Fig. 4 (a) without the door. Fig.

Figure 4 (c) is a front elevation view of the front of the vacuum line cleaning apparatus shown in Figure 4 (a) taken from the center plane of the apparatus.

4 (d) is a side elevational section view of the vacuum line cleaning apparatus shown in Fig. 4 (a) taken from the center plane of the apparatus; Fig.

Figure 4 (e) is a cross-sectional view of the power supply through the connection of the vacuum line cleaning apparatus shown in Figure 4 (a).

4 (f) is a perspective view of the vacuum line cleaning apparatus shown in Fig. 4 (a) including the door.

Fig. 5 is an electric circuit connected to the electrode shown in Fig. 4 (a) and including this electrode. Fig.

6 is a graph showing the effect of an electrostatic collector in one embodiment of the inventive vacuum line cleaning apparatus for particles generated by a typical silicon nitride deposition step.

7 is a graph showing the effect of electrostatic force, attractive force and heat transfer force on the neutral traction force of particles in an embodiment of the vacuum line cleaning apparatus of the present invention.

8 is a micrograph showing the residue accumulation on a silicon piece in a vacuum foreline after a 15 second silicon nitride deposition process.

Figure 9 is a micrograph showing the basic grain size of one grain of the residue shown in Figure 8;

Figure 10 is a micrograph showing the size of the suspended solids deposited on a silicon piece in a vacuum foreline during an experiment performed prior to testing the present invention.

11 is a side cross-sectional view of a second embodiment of the vacuum line cleaning apparatus of the present invention.

12 is a side cross-sectional view of a third embodiment of a vacuum line cleaning apparatus of the present invention.

13 (a) is a side cross-sectional view of a fourth embodiment of the vacuum line cleaning apparatus of the present invention.

Figures 13 (b) and 13 (c) illustrate the effect of the electrostatic traps used in the apparatus of Figure 13 (a) for particles discharged into the apparatus of Figure 13 (a);

Figure 14 (a) is a side cross-sectional view of another embodiment of the vacuum line cleaning apparatus of the present invention.

14 (b) is a view showing the surface area ratio of one electrode to another electrode in the embodiment of the vacuum line cleaning apparatus shown in Fig. 14 (a). Fig.

FIG. 15 is a block diagram of an embodiment of the vacuum line cleaning apparatus shown in FIG. 14 (a). FIG.

16 (a) is a side sectional view of another embodiment of the vacuum line cleaning apparatus of the present invention.

16 (b) is a perspective view of the vacuum line cleaning apparatus shown in Fig. 16 (a). Fig.

Figure 17 (a) is a side cross-sectional view of one embodiment of a vacuum line cleaning apparatus of the present invention that uses microwave power to form a plasma.

17 (b) is a front elevational view of the vacuum line cleaning apparatus of Fig. 17 (a). Fig.

Figs. 18 (a), 18 (b) and 18 (c) are graphs showing voltage waveforms generated by microwave power supply in the apparatus of Fig. 17 (a).

Figure 19 (a) is a side cross-sectional view of a second embodiment of a vacuum line cleaning apparatus of the present invention using microwave power to form a plasma.

19 (b) is a front elevational view of the vacuum line cleaning apparatus of Fig. 19 (a). Fig.

20 is a side cross-sectional view of a basic form of an embodiment of a vacuum line cleaning apparatus of the present invention used in performing an inspection to evaluate the effectiveness of the present invention.

21 is a side cross-sectional view of one embodiment of an apparatus of the present invention optimized for reducing PFC;

22 is a side cross-sectional view of a second embodiment of a PFC reducing device of the present invention.

23 is a side cross-sectional view of a third embodiment of the PFC reducing device of the present invention.

24 is a side cross-sectional view of a fourth preferred embodiment of the PFC reducing device of the present invention.

25 is a side cross-sectional view of a fifth embodiment of the PFC reducing device of the present invention.

26 is a side cross-sectional view of a sixth embodiment of the PFC reducing device of the present invention.

27 is a side sectional view of a seventh embodiment of the PFC reducing device of the present invention.

28 (a) is a side perspective sectional view of one embodiment of a PFC reducing device of the present invention using a gas passing module used in another embodiment of the PFC reducing device of the present invention.

Figure 28 (b) is a side perspective sectional view of one embodiment of a PFC reducing device of the present invention using the gas cylinder and module design of Figure 28 (a).

29 (a) is a graph showing mass spectrum data obtained after forming plasma from CF ₄ and N ₂ O cleaning gas.

FIG. 29 (b) is a graph showing the emission peaks of a specific gas measured during an experiment to examine an embodiment of the present invention. FIG.

Description of the Related Art [0002]

10: chamber 25: RF power source

31: Foreline 32: Vacuum pump system

40: Plasma cleaning device 104: RF matching unit

106: DC / RF filter 454: thin plate anode

468: housing

The present invention achieves this object during processing. That is, in a preferred embodiment, the operation of the present invention appropriately reduces the PFC emissions without requiring additional processing time to prevent the floating solids from accumulating in the foreline. Further, in some preferred embodiments, the present invention does not require additional gas and / or consumables.

In one embodiment of the device according to the invention optimized for reducing particles, a pair of capacitive coupling electrodes define a labyrinth gas path disposed between the inlet and the outlet of the apparatus. Powder residues and other suspended solids that are otherwise collected in the vacuum line when exhausted from the substrate processing chamber (e.g., during the CVD step) are trapped in the gas passages. The apparatus may include a plasma generation system that supplies power to the electrodes to form a plasma within the gas passageway. The plasma is formed from non-reactive discharge gas pumped through the gas path during the cleaning cycle. The components from the plasma react with the trapped suspended solids to convert the suspended solids traps trapped through the discharge line and from the discharge line to gas products that are easily pumped.

In another embodiment, an apparatus of the present invention includes first and second members having opposing surfaces defining a fluid conduit. The fluid conduit includes an inlet, an outlet, and a collection chamber between the inlet and outlet configured and arranged to collect suspended solids flowing through the fluid conduit and prevent the oleophilic material from escaping the collection chamber. A microwave plasma generation system is effectively connected to the apparatus to form a plasma from the etchant gas in the fluid conduit. The components from the plasma react with the suspended solids collected in the collection chamber to form a gaseous product that can be pumped from the fluid conduit. In a preferred embodiment of the apparatus of the present invention, the first and second members are respective electrodes, and the apparatus comprises a particle trapping system for supplying a voltage between two electrodes to collect floating solids on the electrode surface . The plasma reacts with the suspended solids collected electrically to convert the suspended solids into gaseous products that can be pumped from the apparatus.

The gas passageway includes at least one collection chamber configured and arranged to collect the floating solids flowing through the passageway and to prevent the floating solids from escaping the collection chamber. Moreover, the voltage is supplied to at least one electrode to create a voltage between the electrodes that helps to collect and trap the floating solids flowing through the passages.

In another embodiment, the present invention is designed to reduce PFC emissions from semiconductor processing devices. Such an apparatus includes a vessel chamber defining a fluid conduit. The PFC oxidant is in the fluid conduit and the plasma generation system forms a plasma from the emitted PFC gas pumped through the device. The components from the plasma react with the PFC oxidant to convert the effluent PFC gas into less harmful water soluble non-PFC gas products and byproducts.

The apparatus of the present invention provides a PFC oxidant in a silicon containing filter. The plasma generation system produces a plasma from the emitted PFC gas pumped from the apparatus. The components from the plasma react with the silicon-containing compounds in the filter and convert the released PFC gas into less harmful non-PFC gas products and byproducts. In a preferred embodiment of the present invention, the silicon containing compound is a silicon oxide material.

In another embodiment of the present invention, the gaseous silicon source and / or the source of oxygen are introduced into the apparatus to provide a PFC oxidant. The plasma is formed from a gaseous silicon source and / or an oxygen source and a PFC gas. The components from the plasma react to convert less harmful non-PFC gas products and by-products.

In another embodiment of the present invention, the particle trapping and collecting system reduces particles accumulated in the discharge line connected to the substrate processing chamber, and the collected particles and residues provide a PFC oxidant. The particle trapping and collecting system traps silicon-containing residues from a deposition process that produces such residues. The plasma generation system forms a plasma from the emitted PFC gas. The components from the plasma react with the collected residues to convert the released PFC gas into less harmful non-PFC gas products and by-products.

In a preferred embodiment of the present invention, a pair of capacitive connecting electrodes define a labyrinth gas passageway. A DC or AC voltage is applied to the electrodes to create an electric field within the passageway. The electric field pulls the negative voltage particles discharged through the passages onto the positive electrode, and attracts the positively charged particles onto the other electrode. The confined passageway includes at least one region (collection chamber) in which the graft acts to trap the suspended solids discharged through the passageway. The PFC gas discharged through the passage is affected by the RF power supplied to the electrode and is excited into a plasma state. The components from the plasma react with the silicon residue particles trapped within the collection chamber to convert the PFC gas to a non-PFC gas byproduct.

The present invention will be described in detail below with reference to the accompanying drawings.

I. Typical Semiconductor Processing Chamber

The apparatus of the present invention can be used in connection with various other semiconductor processing apparatuses. One suitable apparatus, chemical vapor deposition apparatus, is shown in FIG. 1, which is a cross-sectional view of a simplified parallel plate chemical vapor deposition system 10. The system 10 includes a hydrolysis manifold 11 that distributes the deposition gas to a wafer (not shown) located in the susceptor 12 in the vacuum chamber 15. [ The susceptor 12 responds thermally. The susceptor 12 (and the wafer supported on the upper surface of the susceptor 12) can be controllably moved between the lower loading / off-loading position and the upper processing position 14 adjacent the manifold 11 have.

When the susceptor 12 and the wafer are in the processing position 14 they are surrounded by a bulkhead plate 17 having a plurality of spaced holes 23 that exit into the annular vacuum manifold 24. During processing, the gas inlet to the manifold 11 is evenly distributed across the surface of the wafer as indicated by arrow 21. At this time, the gas is discharged by the vacuum pump system 32 via the vacuum foreline 31 to the circular vacuum manifold 24 through the port 23. Before the gas reaches the manifold 11, the deposition and carrier gases are supplied to the mixing chamber 19 through the gas line 18 where the deposition and carrier gases are connected and then delivered to the manifold 11 .

The controlled plasma is formed adjacent to the wafer by the RF energy supplied from the RF power supply 25 to the manifold 11. The gas distribution manifold 11 is an RF electrode and the susceptor 12 is grounded. The RF power supply 25 can supply one of the single or mixed frequency RF power (or other suitable parameters) to the manifold 11 to improve the decomposition of the reactive species introduced into the chamber 15 .

The circular outer lamp module 26 provides light 27 in the form of an annulus shaped through the quartz window 28 onto the annular outer periphery of the susceptor 12. This heat distribution compensates the intrinsic heat loss pattern of the susceptor and uniformly heats the susceptor and the wafer to achieve deposition.

A motor (not shown) raises and lowers the susceptor 12 between the processing position 14 and the lower wafer loading position. A motor gas supply valve (not shown) connected to the gas line 18 and the RF power supply 25 is controlled by the processor 34 on the control line 36 only partially shown. The processor 34 operates under the control of a computer program stored in a computer readable medium, such as the memory 38. The computer program commands timing, gas mixing, chamber pressure, chamber temperature, RF power level, susceptor portion and other parameters of the particular process.

Typically, some or all of the chamber linings, gas inlet manifold face plates, support fingers 13, and various other reactor hardware are made of a material such as anodized aluminum. Examples of such PECVD devices are described in commonly assigned U. S. Patent No. 5,000,113, entitled " Thermal CVD / PECVD Reactor ", and " Use of In-Situ Multi-step Planarizing Process "

The above-described reactor has been used for explanation, and other CVD apparatuses such as electron cyclotron resonance (ECR) plasma CVD apparatus and inductively coupled RF high density plasma CVD apparatus and the like can be used in the present invention. Also in the present invention, a thermal CVD apparatus, a plasma etching apparatus, a physical vapor deposition apparatus, and other substrate processing apparatus can be used. The apparatus of the present invention and the method for preventing deposition accumulation in a vacuum line are not limited to any particular semiconductor processing apparatus and any particular deposition or etching process and method.

II. The use of a typical invention

During the semiconductor processing process, such as the chemical vapor deposition process performed by the CVD reactor 10, various gas waste products and contaminants are discharged from the vacuum chamber 15 into the vacuum line 31. Depending on the particular process performed, these effluent products include one or both of a suspended solids or PFC gas, such as partially reacted products and by-products, which leave a residue or similar powder material in the foreline when it is discharged through the frontal line can do. The present invention prevents the accumulation of suspended solids in the foreline and reduces the PFC gas emitted from the vacuum chamber (15). Other embodiments of the present invention are designed and used to prevent accumulation of particles or to reduce PFC gas emissions. In addition, some embodiments of the present invention are optimized to reduce particle accumulation and PFC emissions.

As shown in Figure 2, which is a cross-sectional view of the simplified CVD apparatus of Figure 1 suitable for the apparatus of the present invention, the apparatus of the present invention is located downstream of the exhaust gas source. The device may be connected to a portion of the vacuum foreline or may replace a portion of the vacuum foreline. 2, a downstream plasma cleaning apparatus 40 (hereinafter referred to as DPA 40 or DPA) is used along the vacuum line 31 between the vacuum pump system 32 and the vacuum manifold 24. Because of its position, the gas discharged from the vacuum chamber 15 passes through the DPA 40. [ The DPA 40 may be positioned at any location along the vacuum line 31 and the DPA 40 may be positioned in the DPA 40 before the gas exiting the chamber 15 passes through any portion of the vacuum line 31. [ 40 as close as possible to the discharge manifold 24.

It is also possible to connect two or more DPAs to the vacuum line 31, as shown in Fig. Such a structure may use two DPA collecting particles to further protect the vacuum pump 32 from particle and residue accumulation. In the structure shown in Fig. 3, the second DPA 42 is disposed downstream of the DPA 40 in front of the pump 32. In Fig. If any suspended solids leak from the DPA 40, the suspended solids can be trapped in the DPA 42 and converted to gaseous form. Alternatively, the DPAs 40, 42 may be driven by separate RF power sources, respectively, and may be all driven from a main RF power source connected to the processing chamber 10.

These two DPA structures can be used for the two DPAs used for PFC reduction to further limit the emitted PFC gas. In addition, the two DPA structures may include one DPA used for particle reduction and one DPA used for PFC reduction. If individual DPAs used for PFC reduction and particle collection are used, it is desirable to set the position of the DPA used for particle collection downstream of the front of the PFC reduction DPA. This structure can further prevent particle accumulation over the entire foreline and reduce unwanted particle accumulation of the PFC reduced DPA over the next paragraph.

Various structures and embodiments of the DDPA 40 used to reduce particle accumulation of the foreline and reduce PFC emissions will be described in detail below. These embodiments are described for exemplary purposes. In any event, the present invention is limited to these specific structures or embodiments.

A. Specific embodiments of the DPA 40 used for particle reduction

Some embodiments of the present invention are configured and used to reduce particle and residue buildup in the foreline when the particles are discharged from the chamber. As exemplary of such suspended solids for example, as a precursor silane (SiH _4), nitrogen (N ₂₎ and for depositing the silicon nitride using ammonia (NH _3), Si _x N _y H _z, Si _x H _y The residue in the form of brown powder and elemental silicon were observed in the foreline. This residue accumulation occurs from half reaction by-products during the reaction of SiH ₄ + N ₂ + NH ₃ . The silane-based nitrided silicon CVD deposition operation is between substrate processing operations that generate most of the particles. Other substrate processing operations can generate particle accumulation and residue. For example, similar residues are formed during deposition of the silicon nitride layer using disilane (Si ₂ H ₆ ) or other precursor gases or liquids such as organic sources. Residue accumulation may occur during deposition of oxynitride, silicon oxide, silicon carbide, and amorphous silicon films between other layers and may occur during plasma etching and other processing steps.

The particle reduction embodiment of the present invention is a method of trapping suspended solids in a collection chamber to excite residual gases and suspended solids in the collection chamber and reactive gas discharged through the vacuum foreline into a plasma state to accumulate residues and suspended solids Lt; / RTI > The plasma reacts with the residues trapped in the collection chamber and the suspended solids to form gaseous products and by-products that can be pumped through the DPA and vacuum lines without forming a deposit in the line.

During operation, as the deposition gas exits the vacuum chamber through the vacuum line 31, the suspended solids and residues from the gas are mounted on the inner surface of the gas passage in the DPA. Removal of the suspended solids and residues may be performed by operating the DPA 40 to form a plasma within the DPA. The DPA is operated to form a plasma during the cleaning cycle when the etchant gas is evacuated from the chamber 15.

In operation, the DPA 40 creates an electric field that forms a plasma from the exhaust (etchant) gas flowing through the DPA in a plasma state. The plasma promotes the decomposition of the suspended solids in the DPA 40 as gaseous products and by-products that can be pumped through the front surface, thereby preventing particle deposition or residue buildup in the foreline. For example, the residue accumulation in the DPA 40 is in the form of a brown powder containing Si _x N _y H _z , Si _x H _y and elemental silicon for silicon nitride deposition, and the etchant gas used during the cleaning is CF ₄ And a mixture of N ₂ O, the plasma formed by the DPA 40 is decomposed into gas components such as SiO _x , COF ₂ , F ₂ , SiOF ₂ , CO and CO ₂ , NO and O ₂ .

In some applications, the DPA 40 actually holds the plasma formed in the substrate processing chamber, rather than creating a plasma from the etchant gas exhausted into the DPA 40. That is, in these applications, some or all of the plasma formed in the chamber act downstream of the chamber. This may occur during chamber cleaning operations when the plasma is formed from highly reactive fluorinated species. The components from the plasma may be discharged from the chamber to the foreline and DPA in an excited or plasma state. Therefore, in these embodiments, the electric field of the DPA 40 can sustain the plasma. The design and operation of the DPA need not be corrected depending on whether the plasma is maintained in the DPA or generated in the DPA.

It is possible to maintain the plasma during the deposition and cleaning cycles to react further with the CVD gas while the DPA 40 is operating to form and / or maintain the plasma only during the cleaning cycle. In this structure, the additional etchant gas may be introduced into the DPA and into the DPA during the deposition cycle as described in more detail below.

In addition to collecting residues by conventional deposition in the DPA 40, various preferred embodiments of the DPA 40 include a process for removing suspended solids from the chamber 15 in the DPA such that the suspended solids can not be deposited downstream of the DPA. It is designed for trapping oily solids. Traps are made by mechanical, electrostatic and / or thermal trap mechanisms as described in more detail below. Once trapped, the suspended solids are maintained in the DPA 40 until they react with the active species of the plasma during the cleaning process to form gaseous byproducts pumped through the vacuum line 31.

In these embodiments, it is possible to reduce the particle accumulation in the DPA without applying an electric field to form or hold the plasma. This is possible when the ionization of the etchant (fluoro) is sufficiently high during chamber cleaning to have a lifetime long enough to maintain the excited state of the free radicals generated in the cleaning plasma as they exit into the DPA. In the excited state, the free radicals react with trapped suspended solids to convert suspended solids into gaseous products as described above.

Plasma can be generated in the DPA 40 using a variety of known techniques, such as applying HF or RF power to the capacitive coupling electrode or the inductive coupling coil, and microwave or ECR techniques. Some specific embodiments of these methods will be described in more detail below. In each of the embodiments described below, it is desirable that the DPA is designed for low cost. That is, the DPA 40 is preferably designed to prevent particle accumulation in the foreline without requiring the use of any extra cleaning gas or extra cleaning time. It is also desirable that the DPA does not adversely affect film properties such as uniformity, particle contamination, pressure, and the like.

1. Preferred Embodiment

4 (a) -4 (f) are various perspective cross-sectional views of a preferred embodiment of the DPA 40 constructed and optimized to reduce residue and particle buildup. 4 (a) is a front perspective view of the DPA 40 with the door removed. 4 (b) is a front plan view of the DPA (door is removed). 4 (c) is a front perspective cross-sectional view taken along the plane at the center of the DPA. Figure 4 (d) is a side cross-sectional view taken along the plane at the center of the DPA. 4 (e) is a cross-sectional view for supplying electric power through the connection portion of the DPA 40. Fig. 4 (f) is a perspective view of DPA 40 with attached door and handle.

As shown in Figures 4 (a) - (f), the DPA 40 includes an inlet 50 and an outlet 52 (shown in Figure 4 (c)). Between the inlet 50 and the outlet there is a fluid conduit 54 (gas conduit) defined by a pair of opposed aluminum electrodes, namely a cathode 56 and an anode 58. The DAP 40 is connected to the foreline (or directly connected to the process chamber) via the engagement mechanisms 64 and 66 (Fig. 4 (a)). For example, the DPA 40 is connected directly to the chamber outlet by a coupling mechanism 64, and the beginning of the foreline is connected to the DPA at the coupling mechanism 66. The gases and suspended solids discharged from the substrate processing chamber to the foreline travel through the inlet 50 to the DPA 40 and exit through the outlet.

The movable aluminum door 63 (Fig. 4 (d)) seals the gas passage 54 with the back plate 65 (Fig. 4 (d)). The aluminum door 63 and the back plate 65 are electrically connected to the electrode (anode) 58. Electrodes 56 and 58, door 63 and backplate 65 form a sealed vacuum chamber (fluid conduit 54) that prevents leakage of the gas exhaled into DPA 40. The door 63 and the backing plate 65 are connected to a ceramic insulating plate 71 which contacts the electrodes 56 and 58 to form a seal that prevents gas discharged through the DDPA from moving out of the gas flow path indicated by the arrow 60 (Fig. 44 (d)). In a preferred embodiment, a Teflon cushion 73 (Fig. 4 (d)) is included in the door 63 between the aluminum door and the ceramic insulating layer 71. The Teflon cushion 73 has a higher thermal expansion than the ceramic insulating layer 71 and is relatively soft, which allows it to expand without breaking. When the DPA 40 is operated to generate plasma, heat is generated which expands the Teflon layer 73 and applies pressure to the electrodes 56, 58 against the ceramic insulating layer 71. This helps to seal the door 63 so that gas does not leak from the DPA.

The door 63 can be removed with the handle 67 by removing the screw (Fig. 4 (f)) and attached to the DPA 40 (Fig. 4 (f)) via the screw 59. Once removed, the interior of the DPA 40 can be cleaned with a wet solution such as alcohol and / or vacuumed to remove particulate accumulation or residue. In a preferred embodiment, the handle 67 is made of a thermally conductive material such as plastic.

Electrodes 56 and 58 are electrically isolated from each other by four insulating plugs 61 (made of ceramic in the preferred embodiment) referred to as cathode retainers (FIG. 4 (a)). As shown in the figure, the electrodes 56 and 58 have grooves to receive a portion of the cathode retainer. Two cathode retainers 61 are disposed on the front face of the DPA, and the other two cathode retainers are disposed on the rear face of the DPA. In one embodiment, the cathode retainer 61 has a thickness of approximately 1 cm. Thus, the cathode retainer 61 does not extend the entire width of the gas passage 54 and does not block the gas flow through the passage.

For DPA, the gas flows through the fluid conduit 54 as indicated by arrow 60 (Fig. 4 (b)). Fluid conduit 54 includes two mirrorous gas flow paths. The protrusions of the cathode 56 (flow divider-shown in Figure 4 (b)) allow gas to flow into one of the two flow paths. Approximately half of the gas flow is diverted to the left channel of the DPA 40 while the other half is diverted to the right side of the device.

Fluid conduit 54 may include a particulate collection area 62 (see FIG. 4 (a)) that collects and traps particles present in the effluent gas stream, for example, during substrate deposition or other types of processing steps. 0.0 > a)). < / RTI > Each particle collection area 62 is a U-shaped segment of the gas passageway that is arranged to be collected into the lower region of the U-shaped portion by attraction, despite the discharge gas flow path through which the particles are drawn from the DPA. The gas flows through each U-shaped portion by the projecting fingers 79 of one of the cathode 56 or the anode 58 as shown in Fig. 4 (c). These particle collection areas 62 are referred to as attraction or mechanical traps and will be described in more detail below.

Electrodes 56 and 58 form a parallel plate plasma generation system and an electrostatic particle collector. As part of the electrostatic particle trap, DC power is applied to the electrode 56, which is grounded to attract the electrically charged discharged suspended solids. The applied DC power draws the positively charged particles discharged through the DPA onto one electrode and creates an electric field that attracts negatively charged particles to the other electrode. When grounded, the electrode 58 acts as a Faraday cage for RF shielding. As part of the plasma generation system, RF power is applied to the electrode 56. The applied RF power forms a plasma from the effluent gas flowing through the DPA to etch the collected particles and residues within the attraction trap region 62 or along the surface of the electrodes 56,58.

Fig. 5 is a diagram showing an electrical circuit including electrodes 56 and 58. Fig. 5, the electrode 56 is connected to the DC generator 100 and the RF generator 102, and the electrode 589 is grounded. The DC generator 100 supplies the DC voltage required by the electrostatic trap, and the RF generator 102 supplies RF power to form a plasma. The RF matching circuit 104 matches the generator output impedance to 50 OMEGA to minimize the reflected power and the DC / RF filter (low pass RC filter) 106 separates the DC power source 100 from the RF signal interference . The RF generator 102 may be the same power source as the RF power source 25 shown in FIG. 2 and may be a separate RF power source that only derives the DPA 40. Moreover, if multiple processing chambers are present in the wash room, multiple DPAs connected to the chamber can be driven by dedicated dedicated DPA RF power sources connected to an appropriate number of RF power splitters.

For complete reaction of the material flowing through the DPA 40 and deposited in the DPA 40, the DPA is fed to an RF power source (e.g., RF generator 102) at a level sufficient to form and / or hold the plasma Lt; / RTI > Generally, a power level between 50-2000 watts or more can be used depending on the surface area of the cathode and the appropriate intensity of the plasma. The surface area of the cathode 58 is approximately 120in ² according to the embodiment, the power level of between 750-1000 watts (6.31 and 8.42W / in ²⁾ is used. The selected actual power level is determined by balancing the need to use the high power level to form a strong plasma and the requirement to use the low power level to store the energy cost.

The power source driving DPA 40 is operated in a frequency range of about 50 KHz to about 200 MHz or more and preferably in a range of 50 KHz to 60 MHz. Generally, low frequency power supplies are less expensive to obtain and operate than high frequency power supplies. Thus, in most preferred embodiments, the power source driving the DPA 40 is designed to provide an RF frequency of 325 KHz or less. The RF supply power is supplied from either a single frequency RF source or a mixed frequency RF source. The optimum power output and operating frequency of the power supply will depend on the capacity of the gas to be processed in the DPA 40, together with the application and cost problems in which the DPA is used.

Electrical connection to the DPA 40 is made through the power source via the piece (PFD) 68. The PFD 68 is shown in detail in Fig. 4 (e), which is an enlarged side elevational view of the PFD 68. Fig. PFD 68 connects DC generator 100 and RF generator 102 to cathode 56 via connector 70. In the preferred embodiment, the connector 7 is a threaded screw that is connected directly into the cathode 56.

In order to reduce the corrosion of the RF connections and maintain sufficient electrical connection between the screws 70 and the cathode 56, the connection should be made at atmospheric pressure. The region of atmospheric pressure is shown as region 76 and includes the area of the threaded screw 70 that includes the cathode 56. [ The o-ring 78 seals between the cathode 56 and the region 76. A specially designated region is defined by a region of the cathode 56 in which the o-ring 78 is buried from a major portion of the cathode 56, in order to prevent the o-ring 78 from melting by the strong heat that may be generated during operation of the DPA This specially designated region includes the vacuum region 80 and the thin portion 82 of the cathode 56. The region 56A of the cathode 56 may be formed and / And / or the transferred heat is not easily transferred to the region 56B since the vacuum region 80 substantially isolates the cathode region 56B from the cathode region 56A from the region 56B to the region 56A. And a small portion of the cathode 56 carrying the DC signal is thin enough to reduce the heat cut from the region 56A to the region 56B.

The power supply through the connection is received in the aluminum housing 72 and separated from the housing 72 and the door 63 by a Teflon plate 73 and a Teflon ring 74, 75, 81. The housing 72 is electrically connected to the anode 58 and the door 63. The flat washer 84, lock washer 85 and nut 86 assembly allows clamping of the Teflon ring 75 and the Teflon lining 73 to the area 56B of the cathode 56. This clamping force compresses the o-ring 78 to maintain a sufficient seal. The second o-ring o-ring 77 maintains a seal between the teflon lining 73 and the door 63 so that gas does not leak through the power source via the connection 68.

During normal operation, the DC power is supplied to the cathode 56 during the substrate processing step, such as the CVD step, to enhance the particle trapping capability of the DPA 40. The voltage supplied to the electrode 56 varies depending on the application. Typically, applying a voltage between 100-3000 volts creates an effective trap mechanism. This DC voltage can be supplied at all times of the chamber operation (processing and cleaning steps) or can be stopped during the chamber cleaning operation when the DPA 40 is operated.

During substrate processing operations in which silicon nitride is deposited from a process gas of SiN ₄ , N _2, and NH ₃ , approximately 60% ± 10% of the generated particles are positively charged and approximately 40% ± 10% It is determined by experiment that it is charged. It is experimentally determined that the generation of a 500 volts / cm DC voltage in DPA 40, as shown in Figure 6, provides an optimal electrostatic collector for use in substrate processing operations.

In FIG. 6, line 110 represents the total accumulation of negatively charged particles collected on the positively charged electrode by creating an electric field of 200-1200 volts / cm between the electrodes, line 112 being the grounded Lt; / RTI > shows the total accumulation of positively charged particles collected on the electrode. Line 114 represents the total accumulation of trapped particles. At voltages up to 500 volts, the largest particles can not be effectively trapped by the electrostatic collector, and the generation of high voltages creates a particle plasma. This plasma formation changes the characteristics of the generated electric field and reduces the trapping ability.

Electrostatic collectors and mechanical (attraction) trap combinations provide an effective mechanism to prevent evaporation accumulation of the vacuum line 31. The attraction trap is effective when trapping relatively large particles present in the effluent gas stream as the particles are retained in the inner tube 62 by attraction. On the other hand, electrostatic traps are available for very small collecting and trapping particles in the effluent gas stream that are not collected by gravity traps.

As an example, in the deposition of silicon nitride described above, particles with a size of 1 μm and diameters of 1 mm and larger are observed. When these particles are in the discharge line, two important forces are actually acting: gravity (F _g ) and neutral traction (F _nd ). For large floatable solids such as particles with a diameter of 100 μm or more, the main interaction is gravity, and mechanical traps are particularly effective. However, for small particles, the pulling force of the gas may be greater than the attractive force. As a result, the electric field generated between the two electrodes of the electrostatic trap supplies a supplemental force (F _elec ) perpendicular to the trajectory of the particle. This supplemental force (F _elec ) can be greater than the attraction and traction forces for very small particles, such as particles less than 10 μm in diameter, which can lead to very high collection capabilities.

FIG. 7 is a graph illustrating the effect of static and attractive forces compared to the traction force of a particle according to an embodiment of the present invention. Line 122 shows the attraction, line 124 shows the electrostatic force, and line 126 shows the pulling force of the particles. As shown, for small particles, the electrostatic force 124 is greater than the attractive force 122. For large particles, attraction force 122 is greater than electrostatic force 124. In this embodiment, this force causes particles having a diameter of about 30 micrometers to be collected primarily by the electrostatic collector, and particles of about 30 micrometer displacement are collected by a mechanical trap. Regardless of whether electrostatic or gravitational forces modulate a given particle, the focus of FIG. 7 is such that the DPA 40 is designed such that at least one of the electrostatic force 124 or attraction force 122 is greater than the neutral traction force on a given size of particle . In this case, the combination of electrostatic and mechanical trap collectors allows particles of various sizes to be effectively connected.

The thermal force F _th acting on the fourth force acts on the particles in the DPA 40. The heat transfer force is created within the DPA due to the temperature increase and decrease. This temperature increase or decrease can be made by forming in the plasma during the plasma auxiliary cleaning process. During plasma formation, the cathode 56 is hotter than the anode 58 due to ion bombardment and joule effect. In one embodiment, the temperature increase / decrease between the cathode 56 and the anode 58 is 200 DEG C / cm at a gas temperature of 150 DEG. The heat transfer force in this embodiment is shown in FIG. 7 in line 128. If the thermal transfer force 128 is not strong enough to trap particles between 0.1 and 100 micrometers, it will trap the charged and unfilled particles. In addition, those skilled in the art will be able to make the greatest temperature increase so that a large heat transfer force is created to more easily trap particles and residues.

As described above, during the chamber cleaning operation, the RF energy is supplied to the electrode 56 to form and / or maintain a plasma from the released etch gas discharged into the DPA. The components from the plasma react with the particles and residues trapped within the DPA from one or more previous substrate processing apparatuses. Preferably, the supply of RF energy to form a plasma is stopped for a period of time during which the etchant gas is not discharged through the DPA (in this structure, the DPA 40 is referred to as an active element rather than the passive element). Timing control (e.g., switching on and off of the RF power supply 102 and / or the DC power supply 100) of the DPA 40 when the DPA 40 is configured as an active device And is performed by the processor 34 through the application of the control signal transmitted via the control line 36. Although not shown in FIG. 12, this control line is connected to the DPA 40 in the above structure.

In another embodiment, it is desirable to provide the gas supply line directly to the DPA 40 for the introduction of etchant gas separated from the etchant gas in addition to the etchant gas discharged from the chamber 15 during the cleaning operation. This particular gas supply line may be connected to the DPA at or near the inlet 50. It can also be directly connected to the foreline downstream of the DPA. If such an individual gas line is provided, an additional supply of etchant gas may be supplied to the DPA during the deposition sequence or during other substrate processing steps during the cleaning sequence, which may be continuously supplied during the deposition and cleaning cycles. In the embodiment in which the etchant gas is supplied to the DPA during the substrate processing step, the RF energy is supplied to the electrode 56 during the substrate processing step to form a plasma to further etch the deposition material in the DPA.

The effectiveness of the DPA 40 when trapping the particles and reducing deposition buildup depends on a number of factors including the amount of particles generated and discharged from the chamber 15 and the ratio of the emission gas flowing through the DPA 40, The electric field created between the electrodes 56 and 58, the surface area of the electrodes 56 and 58, and the intensity of the plasma generated during the cleaning process.

In addition, a number of other design considerations are to increase the effectiveness of the DPA 40. For example, in the preferred embodiment, the upper surface of the flow divider 57 (Fig. 4 (a)) is curved to a single edge. The experiment shows that the deposition accumulation is collected more quickly at the position where the gas flow contacts the barrier or other surface in the DPA. The sloped surface of the flow divider 57 connected with the inlet gas introduced through the inlet 50 perpendicular to the single edge of the flow divider 57 passes through the inlet 50 for the outlet gas stream input DPA 40 A small contact area is provided to minimize deposition on the upper surface of the flow divider (57). In an experiment performed without an inclined surface (e. G., A rounded surface), particles accumulate on the upper surface of the divider 57. Depending on the amount of this accumulation, the accumulation is broken and falls off the collection area 62. [ If the particle accumulation is large enough, it is not dissipated by the plasma formed during the normal cleaning cycle. This can block the gas path. Also, if the accumulation is an insulating material (e.g. accumulation from silicon nitride deposition), this accumulation interferes with the plasma generation and reduces the intensity of the plasma formed. This can etch the deposited material and block the passageway. Preferably, the side of flow splitter 57 meets an angle of 30 degrees or less to prevent such accumulation. The angle formed may be about 10 degrees and less.

Another design feature that reduces particle accumulation in any one particular region of the DPA 40 is that of the wall 50 for the portion of the gas passageway 54 between the inlet 50 and the point at which the gas flow is distributed into the left and right flows It is outline. In contrast to the inlet with an acute angle, the inlet 50 with an obtuse angle ensures the distribution of the gas flowing into the passageway. This contour transition from inlet 50 to fluid conduit 54 is referred to as a profile manifold.

The gas flowing through the profile manifolds is equally distributed in the flow of gas to the respective left and right portions of the gas passageway 54 to prevent particle accumulation of a portion of the passageway relative to the others. The profile manifold ensures gas distribution across the entire width of the electrode. The contour of the profile manifold is shown in detail in Figures 4 (c) and 4 (d) as the passageway surface 55.

The formation of a uniform plasma can completely remove the particles and residues collected in the DPA 40. The surface of the electrode 56 is approximately the same as the surface area of the electrode 58. Experiments performed in DPA with differences in surface area between electrodes between 3: 1 and 1.3: 1 at various locations are preferred to form a plasma with uneven surface area electrodes, Can be sufficiently removed. However, in this experiment, the accumulation of particles and residues is further removed in the region of the DPA where the electrode surface area ratio is closer to 1.3: 1 than 3: 1. In further experiments in which the surface area of the cathode 56 is within 95 percent of the surface area of the anode (118.79 in ² vs.123.31 in ² ) 58, plasma formation is stronger and particle removal is more efficient. In another embodiment, the surface area of the cathode is the same as the surface area of the anode.

Other plasma uniformity problems include the spacing of the electrodes 56 relative to the electrodes 58. This space is kept constant across the gas passages of the DPA 40 with the following exceptions. Voltage breakdown of the plasma is a function of the pressure and distance between the electrodes (P x D). For the effluent gas stream flowing through the DPA 40, it is necessary that the pressure near the inlet 50 is slightly greater than the pressure near the outlet 52. A larger space is introduced between the electrodes at the bottom of the DPA 40 and the electrodes at the top of the DPA 40 in order to maintain a constant voltage yield in the preferred embodiment. This spatial deformation can be done by making the protruding fingers of one or both of the thick electrode 56 and / or electrode 58 on top of the DPA as shown in Fig. 4 (b). In Fig. 4 (b), the fingers of the cathode 56 and the anode 58 at the top of the DPA 40 have thicknesses a and b, respectively. The corresponding portions at the bottom of the DPA 40 have thicknesses of c and d respectively, where a> c and b> d.

Pressure within the DPA achieves plasma formation. Generally, high pressure causes more effective plasma etching. Thus, operating the DPA at high pressure requires lower power than low pressure, which results in operating cost savings. The high DPA pressure can be obtained by setting the position of the single throttle valve in the foreline after DPA. In such a configuration, it is preferable to use a single throttle valve or preferably a moving throttle valve downstream of the DPA. One throttle valve upstream of the DPA controls the chamber pressure and a throttle valve downstream of the DPA controls the DPA pressure regardless of the pressure in the processing chamber.

Without a throttle valve downstream of the DPA, the pressure in the DPA is equal to the pressure of the foreline (between 0.8-2.5 torr in some PECVD processing units operating at about 4.506 torr). However, in the throttle baffle downstream of the DPA, the pressure in the DPA is controlled over a wide range. Of course, the pressure in the DPA must be less than the pressure in the chamber to maintain the effluent gas stream from the chamber. Increasing the pressure within the DPA has the undesirable side effect of increasing the neutral traction of the particles ejected into the DPA, which reduces the efficiency of the traps. Thus, the actual pressure set in the DPA will balance the particle trap problem with the plasma efficiency problem and will follow the specific application in which the DPA is used.

The pressure sensitive switch 53 (FIG. 4 (d)) may be included to monitor the pressure in the DPA 40. When the pressure within the DPA accumulation is at an undesirable level, the switch 53 sends a signal to the processor 34 to turn off both the DPA and the substrate processing chamber 10. In the preferred embodiment, the switch 53 is a half-standby switch that initiates a shutdown when the pressure in the DPA 40 increases by more than half the air (360 torr).

Depending on the RF power used to form the plasma in the DPA, the size of the cathode, the time period in which the DPA is operated, and other factors, the DPA 40 may include an open cup 69 as shown in Figure 4 (e) . The heat sink pin 69 is attached to the anode 58.

The heat is generated during formation of the plasma at the cathode 56 by the ion collision and Joule effect. As a result, the anode 58 is cooled more than the cathode 56. Alternatively, the anode 68 is thermally insulated from the cathode 56 by a ceramic cathode retainer 61, a ceramic lining plate 71 (at the back and door) and a Teflon insulated ring of the PFD 68. The pin 69 further cools the anode. The fins 69 are made of a thermally conductive material such as aluminum and are a preferred method of cooling DPA 40, where they are passive cooling devices. It is desirable to design the fins 69 so that the exterior of the DPA 40 is cooled to at least 75 占 폚 or below.

In a preferred embodiment in which the DPA is equipped for a DCVD chamber of a P5000 reactor system manufactured by Applied Materials, the fins are disposed on three sides of the DPA, but not on the four axis sides. Instead, four sides (back side) of the DPA are disposed in a portion of the substrate processing chamber. The degree of cooling provided by the pin 69 depends on the size of the fin. In one embodiment in which the temperature of the cathode is between 250-300 캜, the fins 69 are large enough to cool the outside of the DPA to about 75 캜.

Other methods may be used to cool the DPA 40. For example, a cooling system that circulates water around the DPA 40 may be used to transfer heat from the DPA. This cooling system is an active cooling mechanism.

2. Test results using the preferred particle reduction embodiment of DPA 40

In order to demonstrate the effectiveness of the present invention in reducing particle accumulation, a DPA 40 designed according to the preferred embodiment described above is provided for a 6 inch wafer and is attached to a precision 5000 chamber designed for CVD deposition of silicon nitride An experiment is performed. Precision 5000 chambers are manufactured by Applied Materials, Inc., the assignee of the present invention.

Experiments are performed to determine the mixture of residues deposited in the processing chamber by the silicon nitride deposition step after the fluorine cleaning step, before the experiment is run to validate the DPA. The residue mixture is determined for two other silicon nitride deposition / fluoro cleaning process sequences. In each process sequence, the silicon nitride deposition step is the same while the cleaning step is based on the CF ₄ compound in the first sequence and on the NF ₃ compound in the second sequence.

The silicon nitride film is deposited on the wafer by exposing the wafer to a plasma of silane (SiH ₄ ), nitrogen (N ₂ ), and ammonia (NH ₃ ) gases. SiH ₄ flows into the chamber at a flow rate of 275 sccm, N ₂ flows into the chamber at a flow rate of 3700 sccm, and NH ₃ flows into the chamber at a flow rate of 100 sccm. The plasma is formed at a temperature of 400 DEG C and a pressure of 4.5 torr using a 13.56 MHz RF power supply driven at 720 watts. The silicon nitride deposition process continues for approximately 75 seconds to deposit a film of approximately 10,000 angstroms on the wafer.

After the claim for the first sample, the silicon nitride deposition step is completed and the wafer removed from the chamber, the chamber is washed with CF ₄ and N ₂ O for 120 seconds. The ratio CF ₄ to N ₂ O is 3: 1, CF ₄ is flowed at a flow rate of 1500 sccm, and N ₂ O flows at a flow rate of 500 sccm. During the cleaning step, the chamber is maintained at a temperature of 400 DEG C and a pressure of 5 torr. The plasma is formed at a power supply of 13.56 MHz with a power of 1000 watts.

For the second sample, the chamber is cleaned with a plasma formed from NF ₃ , N ₂ O and N ₂ pressure gases. The ratio of NF ₃ to N ₂ O to N ₂ is approximately 5: 2: 10, NF ₃ is introduced at a rate of 500 sccm, N ₂ O is introduced at a rate of 200 sccm, and N ₂ is introduced at a rate of 1000 sccm . The chamber is maintained at a temperature of 400 DEG C and a pressure of 5 torr during the subsequent cleaning process for approximately 95 seconds. The formation of the plasma is achieved with a 13.56 MHz power supply powered by 1000 watts.

The color of the residue for CF ₄ is brown and the color of the residue for NF ₃ washing is yellow white. The residue produced from the Si ₃ N ₄ deposition step is an adaptive color. Thus, these results indicate a complete conversion of the initial brown powder from NF ₃ washing to yellow / white powder. This is demonstrated by the specific free fluoro groups generated in the NF ₃ plasma.

In another series of embodiments, three different residue samples were selected: the powder (Sample A) selected in the foreline at about 0.5 m downstream of the process chamber after the Si ₃ N ₄ deposition step described above, and the NF _{After the 3} / N ₂ O / N ₂ cleaning plasma was run, the powder (Sample B) selected at the same location as Sample A and the inlet of the dry vacuum pump at about 12 m downstream of the chamber after several days of deposition / The selected powder (sample C) is selected. The mixture of powder samples is deduced from hydrogen forward scattering (HFS), X-ray photoelectron spectroscopy (XPS) and X-ray diffraction (XRD) analysis. A mixture of each of these powders is shown in Table 1.

Table 1

Formation of residues from silicon nitride deposition / fluorine cleaning process Sample Residue type Residual mixture C% O% N% Si% F% H% A Brown powder 2.1 41 8.0 33 0.9 15 B Yellow / white powder 0 One 12 8.5 38.5 40 C Orange / brown powder 0.2 6.8 13 42 One 37

Sample A is a direct solid by-product of Si ₃ N ₄ deposition chemistry. The powder reflects the mixture of particles generated in the RF plasma. The powder consists of Si, N, H and oxygen. Oxygen is absorbed from the air during sample collection. Oxygen can not be the initial component of the powder since the oxygen-containing gas is not used during deposition. Similarly, the residual particles generated in the plasma are highly hydrogenated silicon nitride Si _x N _y H _z . This powder is highly reactive. XPS measurements confirm the HFS results and indicate that after the exposure to air, the silicon has 18% elemental, 24% nitride and 58% oxide. Nitrogen has 93% nitride and 7% ammonia. XXRD analysis represents amorphous powder.

Sample (B) is the result of conversion of Powder A after washing using NH ₃ / N ₂ O / N ₂ plasma. The cleaning process completely vaporizes the traces remaining in the deposition chamber, and the conversion is not complete in the foreline due to the limited F ^* free lifetime. However, this lifetime is long enough so that particle conversion can occur at the first meter of the foreline. This white powder has a high F content because Si _x N _y H _z is converted to (NH ₄ ) ₂ SiF ₆ . The polycrystalline white powder exhibits a sublimation temperature of 250 ° C.

The amount of powder accumulated in sample (B) increases, and an increase from the processing chamber indicates that the solid gas is less vaporized and moves less along the foreline. This is due to the dilution of the excited species such as F ^* , CF _X , O ^* between each other during transport from the chamber. In close proximity to the pump, there is a mixture of powders (A, B). This residue (powder C) is brown.

The chemical analysis shows that the powder (C) incompletely transforms from the powder (A). Forming the coated particles, and that could be coated with an initial brown residue block the conversion of the collected powder in the deposition polymer _{- (- CF 2 -C 2 F} 4 -O-) has become concentrated around the x- formed. The accumulation of powder (C) over the foreline of a commercial PECVD silicon nitride system may be greater than 500 grams per month after a continuous deposition / cleaning sequence.

After the mixture of residue accumulation in the chamber is determined, an experiment is performed to determine the grain size of the residue powder. For this experiment, a silicon piece is placed in the foreline to collect the deposited material from the deposition process. Even after the 15 second deposition process, residue accumulation in the form of a brown powder occurs in the vacuum line (3). A micrograph showing this residue accumulation is shown in Fig. The brown powder consists of Si _x N _y H _z , Si _x H _y , SiO _x and basic silicon residues. The basic grain of the residue has a granular shape and a sponge shape with a density of 2.3 g / cm < ³ >. The spherical symmetry of the grain is shown in Fig. 9 and represents growth by even nucleation. Figure 10 is a micrograph illustrating a collection of four or five basic grains (each having a 15-20 micrometer diameter) with a typical residue population having a diameter of approximately 50 micrometers. Further experiments are shown in which the grain size of the powder is increased to the deposition time to form a 1.0 mm or more aggregate during the 90 second deposition step.

The basic DPA is tested to demonstrate the effectiveness of the present invention when reducing particle buildup, such as residue from silane-based nitridation deposition operations. The basic DPA is constructed according to the embodiment shown in Figures 4 (a) - (f) and is connected to the outlet of the P5000 CVD chamber before the foreline. When the experiment, the chamber comprises an operation and CF ₄ / N ₂ O washing step following the three sequential 1.0 micron deposition steps (for three separate wafer) in accordance with the common deposition / clean sequences for the silicon nitride. The deposition / cleaning cycle is continuously repeated for 5000 wafer operation inspection.

For the silicon nitride deposition step, the chamber pressure is set and maintained at 4.5 torr, the chamber temperature is 400 占 폚, and the susceptor is positioned 600 mil from the gas distribution manifold. The deposition gas containing SiH ₄ is introduced at a flow rate of 190 sccm, the deposition gas containing N ₂ is introduced at a flow rate of 1500 sccm, and the deposition gas containing NH ₃ is introduced at a flow rate of 60 sccm. Single frequency RF power with a frequency of 13.56 MHz is supplied at a power level of 455 watts to form a plasma and deposits a layer of silicon nitride at a flow rate of approximately 7500 A / min. The total deposition time for each 1.0 micron layer is about 80 seconds.

During the chamber cleaning step, the chamber pressure is set and maintained at 4.6, the chamber temperature is set at 400 占 폚, and the susceptor (without wafer) is spaced 600 mil from the gas distribution manifold. The cleaning gas containing CF ₄ is introduced at a flow rate of 1500 sccm, and the cleaning gas containing N ₂ O is introduced at a flow rate of 500 sccm. The RF power is supplied to form an etch plasma, which etches the material deposited in the chamber. The frequency power source operates at 13.56 MHz and is driven at 1000 watts. The total cleaning time used to clean the chamber after the three 1.0 micron silicon nitride layer deposition steps is 110 seconds for the first 3000 wafers. The endpoint detector is used to optimize cleaning time for the final 2000 wafers.

The basic DPA has a length of about 35 cm and a diameter of 14 cm. Electrodes 56 and 58 have an overall surface area of 242.1 in ² and are mechanized with aluminum. The cathode is 3.00 inches wide and has a diameter of 39.5966 inches.

A DC voltage of 500 volts is created between the electrode 56 and the electrode 58 for trapping electrically charged particles during the silicon nitride deposition step and the CF ₄ cleaning step, as described above. The electric field is created by supplying 500 volts to the electrode 56 and the ground electrode 58. For plasma formation, the DPA device acts as an active device (i.e., the RF power is provided to the DPA to form a plasma only during the cleaning step, and RF power is not provided during the deposition step). Plasma formation is generated by a 325 KHz RF waveform driven at 1000 watts. The pressure in the DPA is measured at 0.8 torr.

At the time of further testing, the basic DPA is shown to have a 100% effect when preventing particle accumulation in the foreline during a functional inspection of 20,000 wafers using the above-described silicon nitride deposition / CF ₄ cleaning sequence. The use of a basic DPA does not require any additional cleaning gas or any additional cleaning time to trap and remove all suspended solids discharged from the chamber during the experiment. Film properties such as thickness, uniformity, pressure and refractive index for the deposited silicon nitride film during an experiment in which there are no significant changes in any properties between 1 and 5000 wafers (or between any wafers) are measured. Moreover, the number of particles in the chamber is measured during experiments in which the particles with diameters of 0.16 microns or more do not increase during wafer operation.

3. Spiral coil, single tube embodiment

Other embodiments of the DPA 40 incorporating other plasma forming structures are also possible. For example, the plasma is formed by applying an RF signal to an inductive coil, such as a helical reactor coil. The helical coil has a small size and a capacity to form a plasma with a high plasma density. Such coils are known to those skilled in the art and may be designed according to the criteria set forth in many known publications such as Mitchell, et al. Lieberman and Allan J. Lichtenberg, Principles of plasma discharge and treatment of plasma materials. pp. 404-410 John Wiley Sons (1994), incorporated herein by reference.

The helical resonator coil may be made of a highly conductive material such as copper, nickel or gold or similar conductive material. In order to properly resonate the coil, it is very important that the length of the coil is about a quarter or more of the wavelength of the supplied RF signal.

11 is a cross-sectional view of one embodiment of a DPA 40 including such a coil. In Figure 11, the DPA 40 includes a tube 150 that exhausts the gas flowing from the process chamber 15 as it passes through the DPA. The tube 15 is a cylindrical tube made of an insulating material such as ceramic, glass or quartz. In a preferred embodiment, the tube 50 is made of a ceramic material that does not react with the etchant gas, such as fluorine, used in the cleaning step. The tube 150 also has an inner diameter that is the same as the inner diameter of the vacuum line 31. In another embodiment, the tube 150 need not necessarily have a cylindrical shape, but instead may have an inner surface with a flat, oval, or similar curved edge. In these and other embodiments, the inner diameter of the tube 150 may be larger or smaller than the inner diameter of the vacuum line 31.

The coil 152 is wound around the outer periphery of the tube 150 and connected to the RF power supply source at one end of the point 156 and to the ground at the other end of the point 155. The exhaust gas passing through the tube 150 is excited into the plasma state by applying a voltage from the RF power source to the coil 152. In the plasma state, the components from the plasma react with the materials deposited in the tubes to form gaseous products that can be pumped to the DPA 40 and vacuum line 31 by the pump system 32 described above. The coil 152 is a standard spiral resonator coil as described above and can be wound within the tube 150 rather than the outside of the tube.

The outer vessel (154) surrounds the tube (150). The container 154 is used for at least two purposes. First, it acts as a Faraday cage shielded emitter generated by the coil 152. Second, if the ceramic tube 150 is broken or the vacuum seal in the tube 150 is otherwise broken, the vessel 154 provides a second seal to prevent the exhaust gas from leaking. The vessel 154 can be made of various materials such as aluminum or steel or other compounds and is grounded for shielding effectiveness. The upper and lower flanges 157 and 158 connect the DPA 40 to the vacuum manifold 24 and the vacuum line 31 while maintaining a vacuum seal.

The standard RF power supply is designed to have an impedance of 50 ohms as a load. Thus, the coil 152 and the contact point (point 156) to the RF power source should be chosen such that the coil 152 has an impedance of 50 ohms. If the power supply requires a different impedance level, point 156 may be selected.

Coil 152 is driven by an RF power source at a power level of 50 watts or greater. Under these conditions, plasma generation is maximal and uniformity is not critical. The actual voltage generated by the coil 152 follows a number of factors such as the power used by the RF power source, the length of the coil 152 and the winding spacing and coil resistance. Since the voltage is supplied uniformly along the coil, determining the voltage level for the entire coil can be done by determining the level between the point at which the coil is connected to ground and the RF power supply (points 155 and 156). For example, if a particular coil is four times as large as the coil portion between points 155 and 156, then the total voltage of the coil may be four times the voltage level between points 155 and 156. [

The coil, power level and supplied RF frequency should be selected such that a strong plasma is formed in the tube 150 and the voltage generated by the coil 152 does not exceed the current level flowing from the coil to the vessel 154. If the arc is a problem with a particular DPA, it is possible to insert an insulating material between the vessel 154 and the coil 152. However, for simplicity of design, it is desirable to have a space between the air filled container 154 and the coil 152.

The length and size of the DPA 40 may vary. In some applications the DPA 40 may be 4-6 inches and in other applications the DPA 40 may be the entire length of the vacuum line 31 (4-5 feet) can do. It is possible to remove suspended solids more than a shortly designed DPA by selecting long DPA. A short DPA, including an advanced trap mechanism, can collect and trap 99.9% of all suspended solids discharged from the process chamber. Since the length of the coil should be slightly longer than 1/4 of the RF wavelength, there is a correlation between the coil length used and the RF frequency.

The DPA 40 is turned on and off for a certain period of processing, and the DPA can be configured as a passive device. The DPA 40 as a passive device is continuously supplied with sufficient RF power signals so that no specific control signal or processor time is required to turn the DPA on and off.

4. First Spiral Coil, Mechanical and Electrostatic Trap Embodiments

12 is a cross-sectional view of another embodiment of the DPA 40. As shown in Fig. An embodiment of the DPA 40 is shown in FIG. 12, which includes a first inner ceramic tube 160 and a second outer ceramic tube 162. The end of the tube 160 is in the cylinder space of the tube 162 so that gas passes through the DPA 40 as indicated by the arrow 164.

The helical resonator coil 166 is wound around the outer periphery of the tube 162 and is connected to an RF power source 168 as described in connection with the embodiment of FIG. The coil 166 is wound within the interior of the tube 162 or around the outside or inside of the tube 160.

A shell 168 similar to the container 150 described above seals the inner and outer tubes 160, 162. The outer tube 162 may be connected to either the inner tube 160 or the shell 168. In either case, the support structure for the outer tube 162 allows the effluent gas stream to pass through the DPA 40. The support structure may be a flat surface of ceramic material between the tubes 160 and 162 having perforated holes or may constitute 3/4 of the flared connections or fingers extending between the tubes 160 and 162, And can be designed in many other similar ways. The structure including the perforated holes helps to collect and trap the suspended solids within the collection zone 170 described below. However, this structure should be designed such that the holes are large enough not to reduce the flow rate of the gas pumped through the DPA 40.

The preferred design of the DPA 40 facilitates trapping and degradation of the suspended solids. This design is based on the use of a mechanical trap-operated tube (not shown) that collects and holds particles in the exhaust gas stream such that the gas does not travel to the vacuum line 31 through the remainder of the DPA in a manner similar to the trap 62 of Figure 4 And a collecting area 170 of the collecting area 162. These particles are held in the trap and are affected by the plasma until they are broken down or destroyed under the plasma they are formed.

Trap Portion of DPA 40 This operation is dependent on the force acting to hold the suspended solids in the trap despite the discharge gas flow path where it begins to clean the particles through the DPA device into the vacuum line. Thus, the effectiveness of the DPA 40 depends on the ability of the outer tube 162 to prevent particles from remaining in the tube 162 until the particles reach the gaseous product. It is important that the collection area 170 is downstream of the inlet to the DPA and that the DPA 40 is positioned such that the length of the outer tube 162 is sufficient to make the trap with respect to attraction.

Increasing the cross-sectional area of the gas passageway along the plane 176 in the DPA 40 helps trap the floating solids. The flow rate for the effluent gas stream in a given deposition process is generally constant. Thus, increasing the cross-sectional area of the at least one passageway reduces the velocity of the particles in the gas stream which reduces the neutral traction force on the particles. The given particle is trapped by the attractive force in the attraction trap of the DPA 40 if the attractive force on the particle exceeds the neutral traction force.

To further enhance the effectiveness of the mechanical trap, an electrostatic collector 172 may be disposed near the collection region 170. The electrostatic collector 172 may be a small electrode connected to a DC or AC power source. The polarity and amount of charge supplied to the electrostatic collector 172 depend on the polarity type of the suspended solids discharged in each application and the typical charge level.

A variety of other electrostatic trapping devices may be used in the present invention. Other embodiments of such electrostatic collectors will be described in detail below with reference to Figures 13 (a), 13 (b) and 13 (c).

5. Second spiral coil machine and electrostatic trap embodiment

13 (a) is a cross-sectional view of another embodiment of the DPA 40. Fig. The embodiment of Figure 13 (a) uses a mechanical trap design similar to the embodiment of Figure 12 and uses a modified electrostatic trap. Also, the discharge gas is discharged through the side flange 180 disposed adjacent to the upper flange rather than facing the upper flange 181. The flange 180 is disposed for vacuum sealing to the outer casing 184 rather than the outer tube 186. The casing 184 is made of a metal or similar material, and the tube 186 is made of an insulating material such as ceramic.

The RF power is supplied to the DPA through an outer coil 187 that is designed to have an impedance of 50 ohms between the connection point 188 and the point 189 to the RF power source. As described above, the coil 187 should be designed to have an impedance of 50 ohms so that the coil can be driven by a standard RF power source. The inner coil 190 is wound in the inner tube 185. The inner coil 190 receives the RF signal supplied on the outer coil 187 by induction and produces an electric field for driving the plasma reaction.

The center wire 192 passes through the center of the inner tube 185 and the potential is created between the center wire 192 and the inner coil 190 to electrostatically trap the floating solids moving through the DPA. Dislocations can be made using a number of different methods. In each method, the center wire 192 and the coil 190 operate as an electrode. In one embodiment, the center wire 192 is grounded and supplied to both DC or AC coils 190. 13 (b), the particles are attracted by the electric field F _elec produced by the wire 192 and the coil 190, when the emitted particles 194 are negatively charged. Similar results can be achieved if the coil 190 is grounded and a negative voltage is applied to the center wire 192. In this case, however, the wire 192 repels the negatively charged particles toward the coil 190.

In another embodiment, both DC or AC voltages are supplied to the center wire 192, and the coil 190 is connected to the ground potential. In this way, the cathodically charged particles are collected at locations 916 on the positively charged wire 192 as shown in Fig. 13 (c). A similar result can be achieved if a negative voltage is applied to the coil 190 and the center wire 192 is grounded. In this case, the coil 910 repels the negatively charged particles toward the wire 192.

In another embodiment, wire 192 or coil 190 is both not grounded and is both connected to a voltage source that creates a positive or negative potential between wires 192 in relation to coil 190. Of course, in the presence of positively charged suspended solids, this suspended solids can be collected on the electrode opposite to the electrode from which negatively charged suspended solids are collected.

In addition, the particles are collected by electrostatic force in the case where the suspended solids include both positively and negatively charged particles. In this case, positively charged particles are attracted to the low potential electrode and negatively charged particles are attracted to the high potential electrode. It is also possible to supply an AC voltage to the center wire 192. If the AC voltage is connected to the center wire 192 and the coil 190 is grounded, the positive suspended solids are repelled from the wire toward the coil 90 for a positive half cycle. However, during the negative half cycle, negative floating solids are repelled from the wire and collected on the coil 190. In this case, the AC voltage period must be greater than the particle's response time.

In some previous cases, the electric field between the two electrodes may be between 50 and 5000 volts / cm. Preferably, the electric field is between 500 volts / cm (DC) and 1000 volts / cm (AC). The particles are attracted from the center wire 192 to be collected on the coil 190 or are subject to the polarity of the particles and the charge supplied to the coil 190 and the wire 192.

Because this design follows a voltage potential difference created between the coil 190 and the center wire 192, the coil 190 is cooled by the inner tube 1850 to obtain maximum particle collection, The coil 190 and the center wire 192 are in contact with a variety of highly reactive species such as fluorine when placed in the tube 185. Thus, Such as nickel, that does not react with species. Coil 190 creates potentials to attract or repel particles and RF power signals.

6. Third Mechanical and Electrostatic Trap Embodiments Including Parallel Electrodes

14 (a) is a cross-sectional view of another DPA 40 including a mechanical and electrostatic trap. The embodiment of Figure 14 is similar to the embodiment shown in Figures 4 (a) - (c), which forms a plasma from RF power supplied to a pair of capacitive coupling electrodes. However, the electrode in Fig. 14 (a) is the peripheral cylindrical electrodes 402, 404 rather than the parallel plate electrodes of the same surface area in the embodiment of Figs. 4 (a) - (f). As shown in Fig. 14 (b), due to their cylindrical nature, the surface area ratio of electrode 404 to electrode 404 is different in different parts of the DPA. For example, the surface area ratio of the electrode 402 to the electrode 404 is about 3: 1 in the internal passageway 405a. In the same embodiment, the surface area ratio of the electrode 402 to the electrode 404 is about 1.3: 1 in the outer passage 405b.

Electrodes 402 and 404 define a gas passage 405 through which gas exiting the process chamber 15 passes. Electrode 402 is grounded while RF and DC power is being supplied to electrode 404. The RF and DC voltages are supplied to the electrode 404 through the PFD 406. The PFD 406 is insulated from the electrode 402 grounded by the Teflon insulator 408.

The passageway 405 includes a U-shaped attraction trap region 410 having the same shape as the lower half of the circular donut due to the concentric nature of the electrode. The discharge gas is input from the inlet 401 to the gas passage 405 and is output through the outlet 403. [

The DC filter 412 may be coupled to the DPA 40 and the processing chamber 15 so that the voltage supplied to the DPA to trap the electrically charged material in the released gas stream does not interfere with substrate processing operations occurring within the chamber. As shown in FIG.

The electric circuit including the electrode 402 and the electrode 404 is shown in Fig. As shown in Figure 15, electrode 404 is connected to both DC generator 420 and RF generator 422 while electrode 404 is grounded. DC generator 420 supplies the DC voltage required by the electrostatic trap and RF generator 422 supplies RF power to form a plasma. The RF matching circuit 424 matches the generator output impedance to 50 OMEGA to minimize the reflected power and the DC / RF filter (1 m OMEGA resistor in the preferred embodiment) separates the DC power source 420 from the RF power source.

7. Fourth machine with parallel electrodes and electrostatic trap embodiment

16 (a) is a cross-sectional view of another embodiment of a DPA 40 including a machine and an electrostatic trap. The embodiment of Figure 16 (a) includes uniformly spaced parallel electrodes 430, 432 that form a capacitively coupled plasma from the exhaust gas exiting the DPA. Electrode 430 is connected to the RF and DC power sources in a manner similar to the embodiment of Figure 14 (a).

Each electrode is formed of a steel metal and bent to form a gas passage 435. [ The discharge gas from the chamber 15 enters the gas passage through the inlet 434 and is discharged from the outlet 436. The gas passages 435 include two gas flow paths, i.e., paths 435a and 435b, which are disposed side by side and separated by a portion of the electrode 430. [ Dividing the gas flow passage into two separate passages increases the surface area of the electrodes 430 and 432 in the fixed area. The electrodes 430 and 432 should have sufficient thickness within the DPA to prevent them from melting and / or bending under the thermal conditions generated by RF plasma formation. In another embodiment, electrodes 430 and 432 may be mechanized with aluminum.

16 (b) is a perspective view of the DPA shown in Fig. 16 (a). 16 (B), the DPA 40 is surrounded by an aluminum casing 440 similar to other embodiments of the DPA previously described. The casing 440 includes a door 441 attached to the DPA by a screw 442. The DPA 40 can be cleaned by removing this door. In addition, RF and DC power is supplied to the electrode 430 via the PFD connection 438.

8. First Microwave Example

17 (a) is a cross-sectional view of another embodiment of the DPA 40, and FIG. 17 (b) is a front elevational view of the embodiment shown in FIG. 17 (a). The embodiments of Figures 17 (a) and (b) use microwave sources to generate plasma and remove particulate matter and residues collected in the DPA. If a number of different microwave sources are available, a pair of selectively pulsed electron tubes 450 (e.g., electron tubes of the type used in some microwave ovens) are preferably used. Such an electron tube is cheaper than the price of a CW microwave generator or RF generator.

As shown in Fig. 18 (a), each of the electron tubes 450 selectively generates a pulsed (60 Hz) electric field (2.45 GHz). As shown in Fig. 18 (b), by delaying the pulse of one electron tube by a phase difference of 180 degrees with respect to the other electron tube, the two electron tube sources can be pulsed uniformly at 120 Hz as shown in Fig. 18 (c) . In Fig. 18 (c), the first cycle of waveform M1 is generated by one of the electron tubes, and the second cycle M2 is generated by another electron tube. The energy generated by the electron tube can provide ionization efficiency close to 90% at high plasma density. Thus, this power source generates a higher cleaning efficiency than a capacitive coupling electrode that reduces the ionization efficiency between 10-20%.

A further advantage of the microwave source is the heat loss due to the joule effect. Since weak heat is generated, the electrodes 452 and 454 (Fig. 17 (a)) defining the gas flow passage 456 can be easily made thin. The gas flow passage 456 starts from the inlet 458 and terminates at the outlet 460. The passages are divided and follow the dual passages in a manner similar to the gas passages 435 of the embodiment of Figure 16 (a). In addition, inlet 458 flared to DPA 40 at the beginning of gas passageway 456, as shown by profile 464 in Figure 17 (b).

The electron tube 450 is disposed on the opposite side of the DPA 40. Microwave power is connected to the reactor by an appropriate waveguide 462 (Figure 17 (b)). The electron tube and the waveguide are connected to project the microwave through the width of the gas passage 456 so that plasma is generated across the entire gas passage. The spacing between the foil electrodes can be adjusted according to the wavelength of the microwaves so that each node (zero-intensity point) of the electric field is located at the electrode surface (i.e., the spacing between the electrode plates should be a multiple of half a period of the microwave wavelength) . Because of the positioning of the electron tube 450 and the waveguide 462, the plasma is formed in all paths of the gas passage 456. The ceramic gate 466 (Fig. 17 (b)) separates the electron tube and waveguide from the electrode 452, and the outer casing 468 seals the DPA and provides a second level seal.

A DC power source (not shown) is connected to electrode 452 to provide an electrostatic collector as previously described during deposition or other substrate processing operations. The DC power to the electrode 452 is switched OFF (by a switch not shown), and the electrode is grounded when the electron tube 450 is operated during the cleaning operation. The switching electrode 452 is grounded for a period of time required to prevent arcing that may otherwise occur.

9. Second Microwave Example

19 (a) is a side cross-sectional view of another embodiment of the DPA 40, and Fig. 19 (b) is a front elevational view of the embodiment shown in Fig. 19 (a). Figures 19 (a) and 17 (b) are similar to the embodiment of Figure 17 (a) in that it uses an electron tube 450 to generate the DPA plasma .

However, as shown in Figure 19 (a), in the embodiment of the DPA 40, the DPA includes an initial module 472 located downstream of the inlet 474. This module 472 is dedicated to generate plasma during the cleaning cycle such that an etchant (e.g., CF _x and free F when CF ₄ is used as etchant gas) can be generated with increased ionization efficiency. The groups thus generated have a long lifetime and remain activated when they are pumped to the second module 475 of the DPA to react with the deposited and collected material.

The second module 475 includes a gas passage 470 defined by opposing electrodes 476 and 480, which in the preferred embodiment are comprised of a thin plate. The gas passage 470 is similar to the gas passage 456 in the embodiment of Figs. 17 (a) and 17 (b). The gas passage includes the double passages 470a and 470b and terminates at the outlet 478. [

The waveguide 482 is connected to the electron tube 450. The waveguide and the electron tube are arranged so that microwaves form a plasma in the module 472. The inner wall of the anode 476 prevents microwaves from reaching other portions of the gas passageway 470 outside the module 472. Electrode 480 is connected to a DC power source (not shown) to provide an electrostatic collector similar to that previously described. In this embodiment, the DC power for electrode 480 need not be switched OFF during the cleaning cycle. Since the plasma is generated in the second module 475, the arc problem does not occur.

10. Additional particle reduction using a standard DPA Example

In another experiment to demonstrate the effectiveness of the present invention, a second basic DPA 40 is attached to a precision 5000 chamber prepared for an 8 inch wafer. The second basic type DPA is similar to the DPA 40 shown in FIG. 11, except for the design of the bottom flange used to connect the DPA to the foreline. A cross-sectional view of this second basic type DPA and lower flange is shown in Fig. As shown in FIG. 20, the lower flange 200 redirects the exhaust gas flowing through the DPA to the foreline at an angle of approximately 90 degrees. The flange is connected to a quartz window opposite the foreline connection so that the deposited material deposited on the lower portion 204 of the flange can be observed. As described above, the design for the bottom flange of the basic DPA is similar to the U-shaped passageway or mechanical bucket trap design in the embodiment of the DPA 40 shown in Figs. 4 (a) - (f) Lt; RTI ID = 0.0 > 204 < / RTI >

The second basic device includes a quartz tube 206 having a coil 208 made of 3/8 inch copper tubing wrapped around the outer periphery of the quartz tube. The total length of the coil 208 is approximately 25 feet, and the 13.56 MHz power supply is driven at various power levels as described in the description of the experiments below. The quartz tube 206 and the coil 208 are sealed in an aluminum container 210. The total length of the assembly is approximately 14 inches, and the width of the assembly is approximately 4.5 inches.

The effectiveness of the second basic DPA is tested in three separate experiments. For each experiment, 100 wafers were treated in a silicon nitride deposition / CF ₄ fluoro cleaning operation sequence performed in a precision 5000 chamber with a second basic DPA connected between a vacuum discharge manifold and a foreline. The second basic type DPA is kept OFF during the deposition sequence of each experiment and is switched ON by the 13.56 RF power source during the fluoro cleaning sequence. When it is OFF during deposition, the tube is collected along the interior of the tube 206 shown in FIG. 20 as the area 212. These particles are removed from the tube 206 when the DPA is activated during the cleaning sequence. Each of the three experimental conditions is summarized in the following table.

Table 2

Foreline cleaning results Experiment 1 Experiment 2 Experiment 3 RF frequency 13.56 MHz 13.56 MHz 13.56 MHz RF power 200 500 500 CF ₄ flow 1500 2000 2500 N ₂ O flow 500 500 500 result Residue # 1 in Table 2 Residue # 2 in Table 2 The residue removed

In the first experiment, the fluoro wash sequence is 135 seconds and the DPA is driven at 200 watts. CF ₄ enters the process chamber at a flow rate of 1500 sccm and is mixed with N ₂ O introduced into the chamber at a flow rate ratio (3: 1 ratio) of 500 sccm. After a 100 deposition / clean sequence, the DPA is inspected without any residue and deposits. For a flanged flange at the bottom of the DPA, a small amount of residue accumulation is collected. The atomic concentration of this residue accumulation is measured and summarized in Table 3 below. The majority of the silicon in the residue is contained in the form of silicon oxide, with about half of the nitrogen being contained in the silicon nitride film and the other half being in the form of ammonia.

In the second experiment, the fluoro wash sequence is shortened to 120 seconds, and the voltage at which the DPA is driven is increased to 500 watts. CF ₄ flows into the process chamber at a rate of 2000 sccm and is mixed with N ₂ O introduced into the chamber at a flow rate (4: 1) of 500 sccm. After a 100 deposition / clean sequence, DPA is inspected and found without any residues and deposits. A small amount of residue accumulation is collected on the flanges with corners. From the visual inspection, the amount of residue accumulation is about 80% of the accumulation amount in the first embodiment.

The atomic concentration of the residue accumulation is measured and summarized in Table 3 below. As shown in the table, the residue from this experiment contains a higher concentration of fluoro than the residue from the first experiment. The fluoro-enrichment residue provides a further fluoro species for the plasma and makes the residue easier to clean during further DPA activation. In the residue from this experiment, a large number of silicon is contained in the form of silicon oxide, and many of the nitrogen have the form of ammonia.

The third experiment can be completely removed from both the DPA and the flanges with corners where residues are collected during the first and second experiments. In this third experiment, the fluoro wash sequence is extended to 120 seconds, and the voltage at which the DPA is driven is increased to 500 watts. The rate at which CF ₄ enters the process chamber is increased to 2500 sccm and mixed with N ₂ O entering the chamber at a rate of 500 sccm (5: 1). After a 100 deposition / clean sequence, the flanges with DPA and edges are inspected and found without any residue and deposits.

Experimental results and mixtures when residues are present are summarized in Table 3 below.

Table 3

The residues collected at the bottom of the DPA Atomic% Si% N% C O N Si F H elem. nit. ox. Nitride NH ₃ Residue # 1 3.4 44.8 7.4 31.4 13.1 N / A 13.9 20 66.1 48.6 51.4 Residue # 2 4.8 20.5 15.2 19.8 39.8 N / A 4.2 3.3 92.5 3.7 96.3 Residue # 3 none none none none none none none none none none none

B. Specific embodiments of the DPA 40 optimized for PFC reduction

Some embodiments of the present invention are configured and optimized to reduce PFC gas emitted from certain processes that release such gases. The DPA gas thus configured may be referred to as a PFC reduction reactor (hereinafter referred to as PR ² ). For convenience and reference, the DPA 40 configured and optimized as a PR ² device is labeled PR ² (240) in the remainder of the present application. It will be appreciated that the PR ² 240 may be coupled to the chamber, such as the DPA 40 shown in FIG.

A silicon oxide film deposition / cleaning process is used as an example of a process used to reduce PFC emissions in the present invention. However, the present invention is not limited to reducing PFC emissions in the following process sequence, but instead is intended to include any process that introduces PFC gas into the chamber 15 and that the PFC gas becomes a byproduct of the process operation formed in the chamber 15 It should be understood that it applies to certain processes. Additionally, the present invention may be used to reduce the release of other materials such as perfluorocarbons (HFC) or similar gases.

In an exemplary deposition / cleaning sequence, a silicon oxide film is deposited over the material from a process gas comprising silane (SiH ₄ ) and nitrogen oxide (N ₂ O) precursor gases. After the deposition is completed, the substrate is removed from the chamber and a chamber cleaning operation is performed to etch and remove unnecessary silicon oxide deposits from the chamber walls. The cleaning operation is performed by colliding a plasma of CF ₄ and N ₂ O.

As described above, only a small portion of the CF ₄ entering the chamber during the cleaning operation actually reacts with the material deposited on the chamber walls. The remaining unreacted CF ₄ is discharged from the chamber through the foreline with other gaseous components, products and reaction byproducts.

In this example, the PR ² of the present invention forms a plasma from the discharged CF ₄ . The component from the plasma reacts with a silicon source such as solid silicon oxide in PR ² to convert CF ₄ to less harmful products and by-products that do not have a potential damage effect of the PFC. Some of the reactions that generate PR ² inside are:

CF _x + SiO ₂ ----------- SiF _x + CO ₂

CF ₄ + O ₂ ------------ CO ₂ + 2F ₂

2CF ₄ + O ₂ ------------- 2COF ₂ + 2F ₂

C + SiO ₂ --------------- CO + SiO

SiO ₂ + F ₂ --------------- SiOF ₂

Of course, the precise reaction and reaction sequences have become more complicated by the occurrence of basic reactions such as electron collision dissociation of species and reconnection above the gas. None of the above products or by-products, known to be released from PR ² , are PFCs. In fact, each of the above-described products and by-products is water-soluble. Therefore, all the released gases emitted from the PR ² of the present invention are PFC-free gases when the PFC conversion reaction is 100% efficient.

Although the above description is for illustrative purposes only, when a gas other than CF ₄ is introduced into the PR ² of the present invention to form a plasma, the plasma of these PFC gases is not PFC, but forms gas products that are less harmful than PFC And reacts with the silicon oxide source.

In operation, when the PFC gas is discharged from the vacuum chamber 15 to the vacuum line 31, the discharged gas passes through the PR ² 40. Within the PR ² 40, the exhausted gas is susceptible to an electric field that generates and / or holds a plasma. The components from the plasma react with PFC oxidizing agents such as silicon and / or oxygen compounds in the PR ² 40 to convert the released PFC into a less harmful gaseous product and byproduct that is not a PFC and can be pumped through the foreline. Generally, plasma formation is performed only during the cleaning operation of the deposition / cleaning sequence (PR ² being operated), because the PFC gas is present in the discharge stream during this period. Therefore, during the deposition sequence, the plasma is generally not formed in the PR ² (40). However, if PFC gas is released from a special process during deposition or other process steps (e.g., an etch step that reacts with a carbon containing photoresist to produce a CF ₄ byproduct), then PR ² 240 will remove PFC emissions Lt; / RTI >

The silicon and oxygen in the PR ² (240) reacting with the PFC plasma can be of various different sources. In various embodiments of the present invention, the specially configured filter comprises a silicon oxide compound in a solid state such as sand or quartz for plasma reaction. The silicon filter is located in the region of the PR ² 240 where the plasma is formed. In another embodiment of the present invention, the silicon nitride, silicon oxynitride, silicon carbide or similar deposition phase of the deposition / cleaning process sequence is trapped and deposited in a manner analogous to that described with reference to DPA 40 In the PR ² 240. [ Typical residual product which may be trapped during these deposition processes include SiO _2, SiN, SiON, SiC and the like compounds. Of course, the actual residue will depend on the gas introduced during deposition or other process steps. The trap is made using a mechanical and / or electrostatic trap mechanism in conjunction with the heat transfer force as described above in connection with the first aspect and will also be described in detail below with respect to various embodiments of the present invention. Once trapped, silicon residues or other suspended solids remain in PR ² 240 until they react with the active species in the PFC plasma to form a gas byproduct that is pumped through the vacuum line 31. In another embodiment of the present invention, the silicon-containing and / or oxygen-containing gas is introduced into the PR ² 240 specifically to enhance decomposition of the emitted PFC gas. The introduction of these silicon and / or oxygen containing gases may be in addition to or in place of silicon filters and / or electrostatic and / or mechanical trap mechanisms.

The electric field formed in the PR ² 240 to form the plasma is applied to the capacitively connected electrodes by RF power (in some embodiments, HF power (350 KHz) is used instead of RF power to minimize equipment and operating costs) Or < RTI ID = 0.0 > a < / RTI > PFC conversion is directly related to the density of the plasma formed, but in some embodiments such an apparatus that forms a high density plasma, such as an induction coil or a hollow cathode reactor, is preferred. The PFC conversion is directly related to the power associated with the plasma being formed and the residence time of the PFC gas in the PR ² device. Thus, the actual power output of the power source will depend on the application in which PR ² , plasma density, the volume of the PFC gas to be treated in PR ² (240), and the remaining time of the PFC gas among other factors are used. Ideally, PR ² 240 generates a plasma sufficient to convert substantially all of the PFC gas passing through PR ² to another gas.

PR ² 240 should be designed to be easy to use during processing. That is, if PR ² 240 is operated during the cleaning sequence, PR ² 240 is designed to convert substantially all PFC gas from the exhaust stream to non-PFC gas during the cleaning sequence without extending the duration of the cleaning sequence . In such a case, PR ² 240 has an adverse effect on wafer throughput.

The RF power may be supplied by a stand-alone RF power supply that is derived from the RF power supply 25 or only drives the PR ² 240. [ In most embodiments, the microwave ECR is an exception, and it is desirable to use low frequency RF power to operate the PR ² 240. The low frequency RF power supply supplying RF power between approximately 50 KHz and 2 MHz is very inexpensive to operate at higher RF frequencies such as 13.56 MHz. Assuming that multiple processing chambers are present in the wash room, multiple PR ² connected to the chamber may all be driven by an independently provided PR ² RF power source connected to the appropriate number of RF power dividers.

The length and size of the PR ² 240 may vary. In some applications, PR ² 240 may be 4-6 inches longer or shorter, and in other applications, PR ² 240 may be the entire length of the vacuum line 31 replacing the line (4-5 feet or Or more). In general, the residence time of each particle will increase with increasing length and volume of PR ² . The PR ² design should balance the space with consideration of the residual collection effect. However, a short or small volume of PR ² comprising a suitably engineered particle trapping mechanism or filter substantially transforms substantially all of the PFC gas discharged from the processing chamber into a less harmful gas that produces a less critical length and volume factor .

Many other embodiments of the present invention can be configured. Some such embodiments are described below by way of example. The present invention is not limited to these specific embodiments.

1. Silicon Charge Filter Example

a) Single tube, spiral resonator

Figure 21 illustrates a cross-sectional view of PR ² (240). In FIG. 21, PR ² 240 includes a tube 250 that exhausts gas from process chamber 15 as it passes through PR ² 240. The tube 250 is a cylindrical tube made of an insulating material such as ceramic, glass, or quartz. In a preferred embodiment, the tube 250 is made of a ceramic material that does not react with an etching gas such as nitrogen used in the cleaning step. The tube 250 also has an inner diameter that is approximately the same as the inner diameter of the vacuum line 31. In other embodiments, the tube 250 need not be cylindrical, but instead may have an inner surface that is monolithic, flat, elliptical, or similar. In these and other embodiments, the inner diameter of the tube 250 may be larger or smaller than the inner diameter of the vacuum line 231.

The filter 251 is in the tube 250. The filter 251 is a porous filter comprising a solid silicon source that can be used to react with PFC gas under plasma conditions to convert the gas to a non-PFC gas. The filter 251 may be a consumable portion that can be inserted into the tube 250 that can be replaced when the silicon compound is exhausted. The silicon source in filter 251 may be any of a number of silicon-containing materials. Preferably, the silicon source is a silicon-containing material such as sand or glass, quartz, flint or agate. Also, the filter is preferably porous which does not significantly affect the pumping speed or the conductance of the foreline.

The use of silicon oxide materials provides both silicon and coral that the PFC plasma can react to. In a preferred embodiment, the broken quartz is used as the silicon source. Breaking quartz increases the total surface area so that silicon reacts better. Additionally, the broken quartz acts as a mechanical filter for the solid residue produced during the deposition process and provides additional silicon material for reaction when such material is consumed from the chamber 15 during processing to produce silicon residues. I can wrap it.

The coil 252 is wound on the outside of the tube 256 and is connected to the RF power supply at a point 256 and to the ground potential at a point 257. The PFC consuming gas passing through the tube 250 is excited into the plasma by application of voltage to the coil 252 from the RF power source. In the plasma state, the components from the consumable material are removed from the solids of the filter 251 to form gas byproducts that are not PFC pumped in the PR ² 240 and vacuum line 31 by the pump system 32, React with the silicon oxide etchant.

The gas supply line 253 may supply additional gas, which is oxygen and / or silicon source, to enhance the PFC conversion reaction. The gases that may be used may include O ₂ , O ₃ , N ₂ O, SiH _4, or the like. Of course, a liquid source such as tetraethylorthosilicate (TEOS) may be vaporized and flowed through line 53. The rate at which additional reaction enhancing gas enters the PR ² 40 is set by the processor 34 controlling the valve 255. Processor 34 is communicatively connected to valve 255 by a control line not shown.

The coil 252 is an induction coil such as a helical resonator coil. Such coils are well known in the art and are described in detail in < RTI ID = 0.0 > Michael A. < / RTI > Riverman and Alan Jay. Littenberg, Plasma Discharge and Material Handling, page. 404-410 John Willley (1994). The helical resonator coil may be made of a high conductivity metal such as copper, nickel, or gold or similar conductive material. In order to properly resonate the coil, it is important that the length of the coil is somewhat longer or about 1/4 of the wavelength of the provided RF signal. Coils of this length form a stronger, higher voltage field and further improve the decomposition of the PFC gas. The coil 252 can be wound inside the tube 250 more than the outside with respect to the tube.

The outer container 254 surrounds the tube 250. The container 254 is used for at least two purposes. First, it acts as a Faraday cage and protects the CVD processing apparatus 10 and other devices from radiation generated by the coil 252. Secondly, if the ceramic tube 250 breaks or cracks, or if the vacuum seal of the tube 250 is broken in some other way, the container 254 provides a second seal that prevents the consuming gas from leaking out. The container 254 is made of various metals such as aluminum or steel or other compounds and is preferably grounded for shielding effectiveness. Upper and lower flanges 259 and 258 connect PR ² 240 to vacuum manifold 24 and vacuum line 31, respectively, while maintaining a vacuum seal.

Standard RF power is designed with an output impedance of 50 ohms. Thus, the RF power source contact point for coil 252 (point 256) should be chosen such that coil 252 has an impedance of 50 ohms. Otherwise, if the power supply requires a different impedance level, then point 256 should be selected accordingly.

The coil 252 is driven by an RF power source at a power level of 50 watts or greater. The actual voltage generated by the coil 252 depends on a number of factors such as the power used by the RF power source, the length of the coil 252 and the winding spacing and the resistance of the coil among other factors. Since the voltage spreads flat along the coil, determining the voltage level for the entire coil can be done by determining the level between the point at which the coil is grounded and grounded at the RF power supply (points 255 and 256). For example, if a particular coil is four times the coil portion between points 255 and 256, then the total voltage of the coil is four times the voltage level between points 255 and 256. [

The coil, the power level, and the RF frequency provided are strong, a plasma is formed in the tube 250, but the voltage generated by the coil 252 should be selected such that the current does not exceed the level that will arise from the coil to the container 254 . It is possible to place insulating material between the container 254 and the coil 252 if the arc is a problem for a particular PR ² . However, for the sake of simplicity, it is desirable to have a space between the container 254 filled with air and the coil 252.

b) Single tube, microwave

22 shows a cross-sectional view of the second embodiment of the PR ² 240. Fig. The embodiment of PR ² 240 shown in FIG. 22 includes many of the same elements in the embodiment shown in FIG. Thus, for convenience, in Figure 22 and the remainder of the present application, the same reference numerals are used to refer to the same elements. Also, for convenience, the new elements of Figure 22 and other figures are described in sufficient detail.

In Figure 22, microwave generator 260 and waveguide 262 are used to generate a high density plasma from the effluent PFC gas entering PR ² 240. The magnet 264 is disposed around the outside of the tube 250 to further energize the gas particles in the tube 250 and improve the plasma shape, such as in an electron cyclotron resonance (ECR) device. The components from the plasma react with the silicon oxidizer in the filter 251 to convert the PFC gas to a non-PFC gas. As in the above embodiment, additional gas is added to the PR ² 240 from the gas line 253 to improve the conversion process.

Although not shown, the embodiment of PR ² 240 shown in FIG. 22 is preferably sealed to an outer case, such as a container 254. The outer case retains the second seal so that the PFC or other gas passing through the tube 250 does not leak from the PR ² 240 in the event of leakage or other defects of the tube 250.

c) helical coil cavity cathode reactor

23 is a sectional view showing the third embodiment of the PR ² 240. FIG. 23, helical resonator coil 266 is disposed within cylindrical metal tube 268 to form a helical coil cavity cathode reactor embodiment of PR ² 240. In FIG. The coil 266 is connected to the HF or RF power source 269 while the tube 268 is grounded. The remaining structure of the PR ² 240 is not shown in FIG. The structure includes, for example, a gas line 253, a valve 255, flanges 258 and 259, a container 254, and the like, and is similar to the PR ² 240 shown in FIG.

When HF or RF power is applied to the coil 266, an inductively coupled plasma is formed in the coil from the RF power provided in the coil, and an inductively coupled plasma is formed between the coil and the tube 268. Because the coil 266 and the tube 268 are dependent on the reactive nitrogen species from the plasma, they are made of a suitable conducting material such as nickel and react with such species. A silicon filter, not shown, may be disposed within and / or around the coil 266 to provide silicon and oxygen material to react for PFC plasma. Additionally, silicon and / or oxygen containing gas may be supplied to the plasma from the gas line 253.

d) Multistage cavity cathode reactor

24 is a cross-sectional view showing the fourth embodiment of the PR ² 240. FIG. 24, the cylindrical gas passage is formed by a cylindrical shaped anode 272, a cathode 274 and an insulating barrier 275. [ The cathode 274 is connected to the HF or RF power supply 269 while the anode 272 is grounded. An isolation barrier 275 insulates the anode 272 from the cathode 274. This modified electrode / cathode structure forms a multistage, co-cathode reactor in which a high density plasma (10 ¹² ions / cm ³ ) can be formed. Each step (anode / cathode pair) of the reactor forms a high-density plasma in the passageway of the region close to the cathode as shown by region 276.

As well as to increase the particle residence time in the PR ² (240), in order to maintain the common cathode structure and a high plasma density, the pressure in passage 270 is not shown, disposed in the foreline of the back PR ² (240) And can be controlled by a separate throttle valve. The control pressure is 100-500 millitorr (base foreline pressure) up to the actual pressure set to maximize PFC conversion and the pressure in the process chamber (4-20 torr for PECVD processing) and 700 torr or more for SACVD or APCVD processing. Lt; / RTI >

DC power over HF or RF power can be supplied to another embodiment cathode 274 of this multistage, co-cathode reactor design. In a preferred embodiment, however, the HF or RF power is supplied because the directional DC current from the DC power source can etch the electrode. When HF or RF power is used, the sputter etch effect is significantly reduced or absent. HF power is used in the most preferred embodiment to reduce the cost of devices and operations.

As in the cavity cathode reactor design of FIG. 23, a silicon filter, unillustrated silicon and / or oxygen containing gas fed from the gas line 253 may be used to improve the PFC conversion process when appropriate. In addition, structures such as the gas line 253, the valve 255, the flanges 258 and 259, and the container 254 are similar to those of the PR ² 240 shown in FIG. 17 and are not shown in FIG.

2. Silicon particle trap

If deposition, etching, or other treatments that occur in the chamber 15 prior to the PFC cleaning sequence occur in the silicon-containing residue, any embodiment of the inventive apparatus can trap and collect residues for use as a silicon source have. Thus, these embodiments do not require a specific design silicon filter, but can be used.

Examples of deposition processes that produce silicon residues include, but are not limited to, TEOS and silane silicon oxide deposition processes and silane nitride silicon deposition processes. In that process, the released silicon residue trapped for PFC conversion reactions include _{SiO 2, SiN, SiON, SiC} , amorphous silicon and similar compounds. Of course, the actual residues collected will depend on the gas introduced in the deposition, etch or other process used.

The effluent residues from these depositions or other processing operations are coil residues that are collected at the chamber walls and eventually cleaned. Thus, in these embodiments, the conversion of the PFC gas to the non-PFC gas results in the same reaction as the reaction occurring in the chamber during the cleaning operation.

a) Single tube, spiral resonator

25 is a cross-sectional view showing a fifth embodiment of the PR ² 240. Fig. The embodiment of PR ² 240 shown in Figure 25 is based on the fact that the silicon source constituted from the PFC plasma reaction is a silicon-containing residue trapped by an electrostatic collector comprising opposing electrodes 280 and 282 than the solid silicone compound in the filter insert Lt; RTI ID = 0.0 > 21 < / RTI > The silicon-containing residue is trapped and collected during deposition or other processing by the voltage potential applied between the electrodes 280 and 282 from the DC power source 284. The applied voltage potential is positively charged by comparing electrode 282 with electrode 280 (or vice versa). As the remaining particles pass through the PR ² 240, the positively charged electrode 280 and negatively charged particles are attached and collected toward the positively charged electrode 282. Depending on the deposition process type and process length used, a few millimeters or more of silicon-containing residues may be made on the electrodes 280 and 282.

After the deposition sequence is complete and the cleaning sequence is started, the plasma is formed from PR ² 240 from the effluent PFC gas in the same manner as described for FIG. The components from the plasma react with the silicon residues collected on electrodes 280 and 282 to form non-PFC products and byproducts. The voltage potential between the electrodes 280 and 282 may be maintained during the cleaning process to ensure that the collected particles along the electrode remain until they react with the PFC plasma. However, the voltage potential is preferably switched off during the cleaning sequence if the cleaning or other gas used for the particular process etches the electrode. The tube 250, disposed within the electrodes 280 and 282, is in contact with various reactive species such as fluoro. Thus, it is important that the electrodes 280 and 282 are made of a suitable conductive material, such as nickel, that does not react with the species.

A variety of other electrostatic trapping devices may be used in various embodiments of the present invention. For example, a negatively charged DC or AC voltage may be applied to electrode 282 rather than a positive DC voltage. In another embodiment, both electrodes 280 and 282 are connected to a voltage source that forms a positive or negative voltage from electrode 280 with respect to electrode 282. The present invention is not limited to any particular electrostatic collection device.

b) Single tube, microwave

26 is a cross-sectional view showing the PR ² 240 according to the sixth embodiment. In Figure 26, the electrodes of opposite polarity (electrodes 286 and 288) are alternately arranged in the cylindrical volume of tube 250 to form the electrostatic collector as described above. Silicon and oxygen containing residues or similar materials are collected on the surface of electrodes 286 and 288 during the deposition sequence.

As in the embodiment of PR ² 240 shown in FIG. 22, the embodiment of FIG. 26 diffuses high-density plasma from PFC gas passing through the apparatus by application of microwave power from microwave generator 260 and waveguide 262 . A magnet 264 is disposed around the outside of the tube 250 as in an electron cyclotron resonance (ECR) device to further energize the gas particles in the tube 250 and improve plasma formation. The components from the plasma react with the silicon and / or oxidation residues collected on the electrodes 286 and 288 to convert the PFC gas to a non-PFC gas. Additional gas is added to PR ² 240 from gas line 253, not shown, to enhance the conversion process.

Also not shown in the embodiment of the PR ² 240 and shown in FIG. 26 is the container 254 or similar casing mechanism that forms the second seal, so that the PTC or other gas passing through the tube 250 is leaked from the tube Or does not leak from PR ² 240 in the case of other defects.

c) inner and outer cylindrical tubes

27 is a cross-sectional view showing the PR ² 240 according to the seventh embodiment. An embodiment of PR ² 240 shown in FIG. 27 includes a first inner ceramic tube 290 and a second outer ceramic tube 292. The end of the tube 290 is in the cylindrical space of the tube 292, as indicated by the arrow 293, the gas flow through the PR ² 240.

The helical resonator coil 294 is wound around the outside of the tube 292 and connected to the RF power supply 269 as described in connection with the embodiment of FIG. The coil 294 is wound within the interior of the tube 92 or outside or around the inner tube 90.

A shell 297 similar to the container 254 seals both the inner and outer tubes 290 and 292. The outer tube 292 can be held by connecting it to the inner tube 290 or the shell 297. In both cases, it is important that the retaining structure for the outer tube 292 allows a fluid stream of gas to pass through the PR ² 240. For this purpose, the retaining structure may be a ceramic material plane between the tubes 290 and 292 having a plurality of through holes, and may consist of only three of the four or four slider connections extending between the tubes 290 and 292 And can be designed in many different and identical ways. The structure including the through holes can collect and trap silicon residues or other specific materials in the collection area 295 to be described below. Those skilled in the art will appreciate that the structure is designed such that holes are sufficient to reduce the flow rate of gas pumped through PR ² 240. [

The design of PR ² 240 improves trapping and collection of the emitted silicon residue or other particles during the deposition step. The design includes a collection area 295 of the tube 292 that acts as a mechanical trap to collect and retain residues and particles in the consumed gas stream. Residues and particles can be retained in the trap and used to react with components of the PFC plasma formed during the cleaning sequence.

Mechanical operation of the trap portion PR ² (240) depends in part on gravitational force works to outlet despite the gas flow path to clean the particles to a vacuum line through a PR ² device and to keep the particulate material in the traps. Thus, in part, the effectiveness of PR ² 240 follows by preventing the particles from leaving the tube 292 until they are reacted to the gaseous product. For this purpose, it is important that the PR ² (240) is disposed in collection region 295 is facing downward in PR ² from the inlet to the length of the outer tube 292, in cooperation with gravity sufficient to form a trap. Increasing the cross-sectional area of the gas passageway along plane AA within PR ² 240 helps trap particles. The flow rate for the effluent gas stream in any given deposition process is generally constant. Thus, increasing the cross-section of one or more passages greatly reduces the neutral attraction of particulates, thereby reducing the particle velocity of the gas stream. If the particles are given a weight of the particles exceeds the neutral force, it is trapped by the gravity force in the trap of PR ² (240).

To further enhance the efficiency of the mechanical trap, an electrostatic collector 272 including electrodes 296 and 298 connected to a DC power supply 284 may be used as described with respect to FIG.

d) Labyrinth with mechanical and electrostatic trapping mechanism

Figure 28 (a) is a cross-sectional view of a gas passageway module 310 used in another wind interspacing embodiment of a PFC reducing device of the present invention. In FIG. 28 (a), a pair of opposite electrodes 320 and 322 form a gas passage (fluid conduit) through which the discharge gas passes from the processing chamber 15. Module 310 includes both electrostatic and mechanical trapping mechanisms to ensure that all particles emitted from chamber 15 are trapped and collected within the module.

The electrostatic trap is formed by applying a DC voltage to one of the electrodes as described above with respect to FIG. In this way, positively charged particles are collected on one electrode and negatively charged particles are collected on the other.

The mechanical trap further collects the silicon particles and residue by partially relying on gravity to collect the particles in the collection area 324. Each collection region 324 includes a U-shaped segment of the gas passages arranged such that despite the outlet gas flow path to clean the particles that the particles through the apparatus PR ² with a vacuum line, and Collection and maintained in the lower region of the segments. Of course, module 310 may be rotated up and down such that resin region 324 is on the opposite side of the module.

During the cleaning sequence, RF power is provided to one of the electrodes to form a capacitively coupled plasma of PFC gas passing through the module. Preferably, the electrodes 320 and 322 are designed to have substantially the same surface area. Such a design allows a uniform plasma to strike through the entire area / passageway formed by the electrodes. As in the above embodiment, the components from the plasma react with the collected silicon particles and residues to convert PFC gas to non-PFC gas.

Electrostatic collectors and mechanical trap bonding provide a particularly effective mechanism for collecting the emissive silicon residue from the chamber 15. In practice, such coupling has the additional advantage of providing a collection effect of almost 100%, thereby eliminating or preventing deposits on the vacuum line 331. As discussed above with respect to FIG. 6 and DPA 40, the mechanical trap section is particularly useful for trapping relatively large particles present in the effluent gas stream since these particles are more likely to remain in collection chamber 324 by gravity effective. On the other hand, electrostatic traps are effective in collecting and trapping smaller particles in the effluent gas stream, but are not collected by mechanical traps. Also, as noted above, thermophoretic forces due to temperature gradients between the electrodes can be used to trap the particles.

The module shown in FIG. 28 (a) can be used as part of various embodiments of PR ² 340. One embodiment of such an embodiment is shown in FIG. 28 (b) and includes a PFC reducing device of the present invention that uses a portion of the gas passage module design of FIG. 28 (a) stacked on top of other similar portions of the module. Sectional side view showing an embodiment. Of course, other designs with modules or like modules shown in Figure 28 (a) are possible. For example, three, four, or more modules may be placed in a sequence to form relatively long gas passages having increased electrode surface area for a particular collection. In addition, three, four or more modules may be stacked on top of each other and connected in a manner similar to the embodiment shown in Figure 28 (b). Module 310 may be equipped with a filter element comprising an additional source of silicon through which components from the PFC plasma may react. The possibilities for other design variations based on the module 310 are almost infinite.

28 (b), the outflow gas from the chamber 15 enters the PR ² 340 through the inlet port 330 and is discharged through the outlet port 332. The divider 334 ensures that the gas flows through the labyrinthal passage formed by the electrodes 320 and 322 along arrow 323. When PR ² (340) is be directed vertically, the inlet 330 is larger than the particle emitted through the passage and on the side surface along the axis (AA) is intended to collect in the collection region 324 under the gravity. If the PR ² 340 is oriented vertically, the inlet 340 is laterally along the axis BB and larger emission particles through the passages are to be collected in the collection area 325.

DC generator 338 supplies electrode 322 with a positive DC voltage during which the electrode 320 is grounded and during the deposition and cleaning sequence. Thus, the negatively charged particles are intended to be collected on the surface of the electrode 322 and the positively charged particles are intended to be collected on the surface of the electrode 320.

As in other embodiments, the RF generator 336 provides RF power to the electrode 322 during the cleaning sequence to form a plasma from the outgoing PFC gas in the passages between the electrodes 320 and 322. The plasma reacts with the collected silicon along the collection region 324 or 325 and the electrodes 320 and 322 to convert the PFC gas to non-PFC gas products and byproducts. The DC / RF filter 340 prevents the RF power source from interfering with the DC generator 338. DC and RF power is applied to electrode 320 rather than electrode 322; However, it is desirable for the electrode 320 to be grounded for safety and radiation.

3. Usage and test results on PFC reduction DPA

To indicate the effectiveness of the present invention, examples are basic (PR ² 340) is affixed to the external equipment the Precision 5000 chamber executed for the 8-inch wafer. The Precision 5000 chamber is manufactured by Applied Materials, the assignee of the present invention.

In the embodiment, basic type PR ² is similar to PR ² 340 of FIG. 21 except that filter 351 is not included in ceramic tube 350 and no additional gas supply line 353 is present. The total length of the PR ² is about 25 inches and the diameter of the tube 350 is about 1.5 inches. PR ² is attached to the chamber lower throttle valve back precision 5000 chamber.

The analysis outlet gas is emitted from the CF ₄ and N ₂ O in three different steps: Measurements are made in front of the vacuum pump below about 20 feeds from the chamber. So, it is considered to be detected only by the stable paper RGA. Because the bulk analysis is complex, the deposition step is not performed before the cleaning sequence.

The experimental conditions are as follows. The pressure in the chamber is maintained at ² torr, resulting in a corresponding pressure of 0.5 torr at PR2. CF ₄ and N ₂ O are introduced into the chamber at a rate of 500 sccm. The plasma formed in the chamber is driven by a 13.56 MHz RF power supply at, 1000 watts (RF1) during which is driven by a 13.56 MHz RF power source at a 900 watts (RF2) formed in the plasma PR ^2.

During the first stage of the embodiment, the cleaning gas enters the chamber and is allowed to flow through the chamber and the PR ² without forming a plasma. In the second step, the plasma is formed within the Precision 5000 chamber, but not within the PR ² . In the third step, a plasma is formed on both the chamber and the PR ² . The results of these embodiments are shown in Fig. 29 (b). The first step determines the spectrum emitted from the process and setting based on the relative analysis of CF ₄ emissions.

29 (a) shows the mass spectrum obtained when the plasma is formed from both the chamber and the PR ² cleansing gas. It is important for the RGA device to detect gas by ionization. Thus, detection of CF ₃ ⁺ , CF ₂ ⁺ and CF ⁺ ions represents effluent CF ₄ . (Indicated in parentheses) in Fig. 29 (a), peaks ^{C + (12), N +} (14), 0 + (16), F + (19), CF + (31), O 2 + (32 ), F ₂ ⁺ (38), N ₂ O ⁺ (44), CF ₂ ⁺ (50) and CF ₃ ⁺ (69). Each peak corresponds to a product during the decomposition of the initial gas reactants (CF ₄ and N ₂ O). The peaks corresponding to CO ⁺ (28), CO ₂ ⁺ (44), COF ⁺ (47), COF ₂ ⁺ (66) and COF ₃ ⁺ (85) correspond to the byproducts of the reactions occurring in the chamber and PR ² . Possible interpretation may be due to CO ₂ ⁺ and N ₂ O ⁺ (line 44). By recording the spectra and nonplasma of one CF ₄ and N ₂ O and the response and determining that the peaks in line 44 exhibit 90% CO ₂ ⁺ and 10% N ₂ O ⁺ when RF 1 and RF ₂ are on do.

Qualitatively, a decrease in CF ₄ is observed when the response of the CF ₃ ⁺ (69), CF ₂ ⁺ (50), and CF ⁺ (31) peaks decreases. Further degradation evidence is observed when the peak response corresponding to N ₂ O decreases. The byproduct response of the reactions CO ⁺ (28), CO ₂ ⁺ (44), COF ⁺ (47), COF ₂ ⁺ (66) and COF ₃ ⁺ (85) increases in proportion to the decrease of CF ₄ .

Figure 29 (b) shows the power generation peaks of the specific gas measured by RGA during each of the three step embodiments. Particularly, FIG. 12 (b) shows the responses of the peaks 44 (N ₂ ⁺ ), 69 (CF ₃ ⁺ ) and 28 (CO ⁺ ). The first 80 seconds shown in Figure 29 (b) shows the response of these gases when no plasma is formed in the chamber or in PR ² . For the next 80 seconds, the plasma is finally formed only in the chamber, plasma is formed in the chamber and PR ² for 160 to 240 seconds.

From Fig. 29 (b), as the plasma bumps into the chamber, the amount of radiated CF ₄ and N ₂ O is reduced and the amount of emitted CO (the main by-product of the CF ₄ conversion process) increases. Activating PR ² (40) (and forming a plasma in PR ² 40) further reduces CF ₄ emission and causes a total CF ₄ reduction of about 30%.

In another embodiment where the results are not shown, a total reduction of about 50% is achieved by increasing the pressure in the PR ² to about ² torr. Thus, the embodiment shows that the device of the present invention preferably reduces PFC. Additional reductions can be made by incorporating one or more additional PR ² features discussed within the present application. Further, since CF ₄ is difficult to convert PFC gas, additional results produce better results for conversion of other PFC gases.

By discussing several different embodiments of the present invention, many equivalent or other apparatus and methods for removing particles from a vacuum line in accordance with the present invention will be apparent to those skilled in the art. In addition, many equivalent or other apparatus and methods for reducing PFC radiation from a process chamber in accordance with the present invention will be apparent to those skilled in the art. Additionally, although the present invention has been described with reference to illustrative embodiments for simplicity and understanding, it is evident that several changes and modifications are made. For example, in one embodiment, the mechanical particle trap of the present invention is described as an inner passageway enclosed by an outer passageway, such a trap is not included in the periphery of the second passageway but instead extends A first passage is formed. In another embodiment, the described gas passages are designed with labyrinthine shapes (and include gravity traps) in a manner similar to the passages illustrated and shown or otherwise illustrated in Figures 28 (a) and 28 (b) do. Silicon particle trapping includes a separate filter element filled with quartz or other silicon containing compound to improve PFC decomposition if the silicon residue collected in the electrode is insufficient. Further, embodiments without a silicon filter and a particle trapping system are possible. In these embodiments, a gas such as SIH ₄ or O ₂ for the PFC conversion reaction flows into PR ² 340 through a gas supply line such as line 253. Additionally, the electron tube used in the embodiment of the DPA 40 shown in Figures 17 (a) and (b) and 19 (a) and (b) forms a plasma in various embodiments of the PR ² 40 . These equivalent and other modifications are within the scope of the present invention as evidenced by obvious changes and modifications.

The present invention has the effect of preventing PFC solids and other residues from accumulating in the discharge line of the substrate processing apparatus and reducing PFC emissions from the substrate processing apparatus.

Claims (17)

An apparatus for minimizing deposits in a discharge line of a substrate processing chamber, A fluid conduit having an inlet, an outlet, and a collection chamber between the inlet and the outlet, the collection chamber having a first portion and a second portion, the first portion and the second portion having an opposing surface defining a fluid conduit, Wherein the oily solids are collected and configured so that the oily solids are not discharged; And a microwave plasma generation system used to form or maintain a plasma from the etchant gas in the fluid conduit. The method of claim 1, wherein the first member comprises a first electrode and the second member comprises a second electrode; Further comprising a particle trapping system operatively connected to the electrodes and having means for supplying a voltage between the electrodes for collecting the electrically charged solids on the surface of the counter electrode. . 3. The apparatus of claim 2, wherein the etchant gas is discharged from the substrate processing chamber through the discharge line during a cleaning operation of the substrate processing chamber. 3. The apparatus of claim 2, wherein at least a portion of the etchant gas flows into the top and bottom of the discharge line from the substrate processing chamber. 3. The apparatus of claim 2, wherein at least a portion of the etchant gas flows directly into the fluid conduit. 3. The apparatus of claim 2, wherein the fluid conduit is at least partially defined by the first and second electrodes and communicating the collection chamber with the inlet and the outlet, respectively, to prevent the emission of suspended solids from the collection chamber Wherein the first and second passageways comprise vertical first and second passageways. 7. The apparatus of claim 6, wherein the fluid conduit formed between the opposing electrode surfaces defines a plurality of collection chambers sequentially formed in the fluid conduit between the inlet and the outlet. 8. The apparatus of claim 7, wherein the counter electrode surface defines an S-shaped fluid conduit. 8. The apparatus of claim 7, wherein the at least partially vertical first and second passageways extend along a length of the fluid conduit; Wherein the fluid conduit has a width that extends between opposite sides of the device. 10. The apparatus of claim 9, wherein the microwave plasma generating system comprises first and second electron tubes. 11. The apparatus of claim 10, wherein the first and second electron tubes selectively generate pulses such that the phase of the second electron tube is retarded by 180 degrees with respect to the phase of the first electron tube. 12. The apparatus of claim 11, wherein the first and second electron tubes are disposed on opposite sides of the apparatus; Further comprising first and second waveguides operatively connected to the first and second electron tubes, respectively. 13. The apparatus of claim 12, wherein the first and second electron tubes and the first and second waveguides are arranged to provide microwaves to the first and second passages at least partially perpendicular to the fluid conduits Device. 14. The apparatus of claim 13, wherein the distance between the first and second electrodes is substantially equal to the wavelength of the microwave generated by the first and second electron tubes. 15. The apparatus of claim 14, further comprising a switch for connecting the first and second electrodes to ground; Wherein a voltage is applied to the first and second electrodes during a substrate processing operation to collect the electrically charged floating solids, the switch being configured to ground the electrode during a substrate cleaning operation when a plasma is formed from the etchant gas, To the second antenna. 12. The apparatus of claim 11, wherein the first and second electron tubes are disposed on the same side of the apparatus; And a waveguide operatively connected to the first and second electron tubes. 17. The apparatus of claim 16, wherein the first and second electron tubes are arranged to provide microwave to the first passageway that is at least partially vertical but not at least partially perpendicular to the second passageway.